Professional Documents
Culture Documents
CASTEP Transition State Search dialog 170 Electron localization function selection 217
CASTEP TS Confirmation dialog 171 Energy evolution selection 217
Electronic tab 172 Fermi surface selection 218
CASTEP Electronic Options dialog 174 IR spectrum selection 218
Properties tab 186 NMR selection 219
Band structure selection 186 Optical properties selection 220
Core level spectroscopy selection 188 Optics Analysis Options dialog 221
Density of states selection 190 Orbitals selection 221
Electron density difference selection 193 Phonon dispersion selection 222
Electron localization function selection 193 Phonon density of states selection 223
Electronic excitations (TD-DFT) CASTEP Phonon DOS Analysis Options
selection 194 dialog 224
NMR selection 194 Population analysis selection 224
Non-linear optical properties selection 196 Potentials selection 225
Optical properties selection 196 Reaction kinetics selection 226
Orbitals selection 199 Raman spectrum selection 226
Phonons selection 199 STM profile selection 227
Polarizability, IR and Raman spectra Structure selection 228
selection 203 Thermodynamic properties selection 228
Polarization selection 205 Results file selector 229
Population analysis selection 205 Chapter 6: CASTEP keywords 230
Solvation energy 206 Appendix I: CASTEP References 231
Stress selection 206 Dassault Systèmes Support Resources 240
Job Control tab 206
CASTEP Job Control Options dialog 208
CASTEP Job Files dialog 209
CASTEP Analysis dialog 210
Band structure selection 210
Core level spectroscopy selection 211
CASTEP EELS Analysis Options dialog 212
Density of states selection 213
CASTEP DOS Analysis Options dialog 214
Elastic constants selection 215
Electron density selection 216
Electron density difference selection 216
Additional Information
For additional information about BIOVIA Materials Studio and other Dassault Systèmes software
applications, visit the Dassault Systèmes Customer Support site:
https://www.3ds.com/support/
Note: CASTEP cannot handle structures containing pi-system objects. Such objects should be
removed from the document prior to running a CASTEP calculation.
Note: CASTEP performance can be greatly increased if the symmetry of the structure is taken into
account. CASTEP offers an option of finding and using the symmetry automatically if the input
structure is described as having a P1 symmetry. This option is particularly relevant for calculations
on molecular crystals and crystal surfaces, since these structures are typically created as P1 (or p1)
systems. Automatic conversion to a higher symmetry structure will honor all atomic properties
that are relevant to the current CASTEP settings, such as constraints and electronic configuration.
Tip: The time required for a CASTEP calculation increases with the cube of the number of atoms in
the system. Therefore, it is recommended that you use the smallest primitive cell description of
your system whenever possible. Select Build | Symmetry | Primitive Cell from the menu bar to
convert to a primitive cell.
Calculation setup: Once a suitable 3D structure document has been defined, then it is necessary to
select the type of calculation to be performed and set the associated parameters. For example, in the
case of a dynamics calculation, these parameters include the ensemble and its settings, the
temperature, and the number and length of the time steps. Finally, the server on which the
calculation is to be run should be selected and the job initiated.
Analysis of the results: When the calculation is complete, the files related to that job are returned to
the client and, where appropriate, displayed in the Project Explorer. Some further processing of
these files may be required to obtain observables such as optical properties. The tools on the CASTEP
Analysis dialog may be used to visualize the results of the calculation.
To select a CASTEP task
1. Choose Modules | CASTEP | Calculation from the menu bar to display the CASTEP Calculation
dialog.
2. Select the Setup tab.
3. Select the required CASTEP task from the Task dropdown list.
CASTEP geometry optimization is based on reducing the magnitude of calculated forces and stresses
until they become smaller than defined convergence tolerances. It is also possible to specify an external
stress tensor to model the behavior of the system under tension, compression, shear, and so on. In
these cases the internal stress tensor is iterated until it becomes equal to the applied external stress.
The process of geometry optimization generally results in a model structure that closely resembles the
real structure.
The accuracy of the lattice parameters calculated using CASTEP is illustrated in Figure 1 (Milman et al.,
2000).
Where:
V0 is the equilibrium volume
A detailed comparative study of various analytical forms of EOS was performed by Cohen et al.(2000).
Note: The values of B and B' obtained from such fitting experiments depend on the pressure range
used in the calculations. Experimental values obtained using a diamond anvil cell are usually in the
range 0-30 GPa, so this is the range recommended for theoretical studies. It is also important to avoid
using negative values of pressure in such studies. In addition the pressure values used to create the
P-V dataset should not be uniform. More accurate sampling of the low-pressure region is required to
obtain an accurate value of the bulk modulus.
Note: Quantum-mechanical molecular dynamics calculations usually require a smaller time step than
forcefield based dynamics applications. However, this is less of an issue in CASTEP because
wavefunction and density extrapolation are used.
When a 3D periodic molecular or crystal structure is built, it usually needs to be refined to bring it to a
stable geometry. The refinement process is known as optimization. It is an iterative procedure in which
the coordinates of the atoms are adjusted so that the energy of the structure is brought to a stationary
point, that is, one in which the forces on the atoms are zero. A transition state is a stationary point that
is an energy maximum in one direction (the direction of the reaction coordinate) and an energy
minimum in all other directions.
During the course of chemical reaction, the total energy naturally changes. Starting from the reactants,
the energy increases to a maximum and then decreases to the energy of the products. The maximum
energy along the reaction pathway is known as the activation energy. The structure corresponding to
this energy is called the transition state.
The TS Search task is particularly useful for predicting barriers to chemical reactions and determining
reaction pathways. It can be also used to find diffusion barriers for either solid state diffusion or surface
diffusion. The reactants and products can either be optimized separately before running a TS Search or
as part of this task. In the latter case partition functions for reactant, product, and transition state can
also be calculated. This information can be used to calculate the reaction rate coefficient.
Use the controls on the CASTEP TS Confirmation dialog to specify the convergence criteria for each
individual geometry optimization (microiteration), and to specify the number of frames to use in the TS
confirmation trajectory.
The output of a TS confirmation calculation is another trajectory document. This one follows the
Intrinsic Reaction Path (IRP) as discussed in Reaction pathways. Each successive refinement to the
pathway is displayed as a graph in Materials Studio. Use the chart selection tools to choose any point on
the graphs and display the corresponding geometry.
In most cases, no minima will be found on the IRP other than your reactant and product. If so, the TS
confirmation calculation verifies that the TS does connect your presumed reactant and product.
On the other hand, any other minima located by the TS confirmation are the structures that are actually
connected by the TS. These are identified with an asterisk (*). You should perform a geometry
optimization to refine the geometry of these minima. Select the structures as described in the Displaying
trajectory and chart data topic.
You can generate a detailed map of the entire reaction by repeating TS Search and TS Confirmation for
all minima on the IRP.
CASTEP Properties
The CASTEP Properties task allows you to compute electronic, structural, and vibrational properties
after the completion of a single-point energy, geometry optimization, or dynamics run on a 3D periodic
system.
Note: To calculate properties using the CASTEP Properties task, the results files from a suitable
simulation must be present in the current project.
You can also use the Properties tab on the CASTEP Calculation dialog to request calculation of such
properties as part of a CASTEP run. You can view the results using the CASTEP Analysis dialog.
The properties that CASTEP can generate are:
Band structure: This calculates electronic eigenvalues along high symmetry directions in the Brillouin
zone. These calculations are non-self-consistent for both valence and conduction bands, using
electronic charge densities and potentials generated during the simulation.
Core level spectroscopy: This calculates electronic energies on the Monkhorst-Pack mesh of k-points
and the matrix elements for electronic interband transitions, either with or without core holes.
Density of states: This calculates electronic eigenvalues on a fine Monkhorst-Pack grid . These
calculate are non-self-consistent for both valence and conduction bands, using electronic charge
densities and potentials generated during the simulation.
Electron density difference: This calculates the electron density difference with respect to a linear
combination of either the atomic densities or the densities of sets of atoms contained in the
structure.
Electron localization function: This calculates a simple measure of electron localization in atomic
and molecular systems.
Electronic excitations (TD-DFT): This calculates electronic excitations for molecules in a box, using
the time-dependent density functional theory (TD-DFT).
NMR: This calculates chemical shielding tensors, electric field gradients, J-couplings, and g-tensors.
Note: NMR in CASTEP is part of the separately licensed module NMR CASTEP. NMR calculations
can only be performed if you have purchased this module.
Optical properties: This calculates matrix elements for electronic interband transitions . The CASTEP
Analysis dialog allows you to generate grid and chart documents containing measurable optical
properties.
Orbitals: This provides information about electronic wavefunctions. This allows you to visualize 3D
distribution of various electronic states (orbitals). Visualization of STM profiles also requires this
information.
Phonons: For phonon dispersion runs, this calculates phonon frequencies and eigenvectors along
high symmetry directions in the Brillouin zone. In the case of phonon density of states calculations,
this computes phonon frequencies and eigenvectors on a Monkhorst-Pack grid. Analysis requires
this information to display total and projected (partial) phonon densities of states. Calculation of
thermodynamic properties including atomic displacement parameters also uses this information.
This allows assignment of temperature factors during analysis.
Note: Phonon calculations take into account existing fixed atom constraints, regardless of the
overall Task setting. Calculation of vibrational properties does not include such fixed atoms,
corresponding to the "partial Hessian" approach.
Tip: Select the Fix fractional position checkbox on the Atom tab of the Edit Constraints dialog,
accessible from the Modify menu.
Polarizability, IR and Raman spectra: This calculates the optical (ω = ∞) and dc (ω = 0) dielectric
permittivity or the optical (ω = ∞) and static (ω = 0) molecular polarizability, along with infrared or
Raman intensities (response to an electric field in the infrared range). Permittivity is relevant for solid
materials, while polarizability and infrared intensities are relevant to molecules prepared using the
supercell approach.
Population analysis: This performs Mulliken analysis and Hirshfeld charge analysis. This includes
calculation of Mulliken bond populations and angular momentum-resolved atomic charges (as well
as magnetic moments for spin-polarized calculations, and production of Hirshfeld atomic charges.
Optionally, CASTEP can generate the weights required for partial density of states (PDOS)
calculations.
Solvation energy: This calculates the free energy of solvation and writes it to a seedname_
Solv.castep file. CASTEP automatically prepares calculations using vacuum and implicit solvation
schemes and extracts solvation energy from the respective energies of these two calculations.
Stress: This calculates the stress tensor and writes it to a seedname.castep file. This information
is useful if, for example, you perform a geometry optimization run with fixed cell parameters and you
want to verify how far the lattice is from equilibrium. For example, carry out a supercell study of a
point defect with the fixed cell that corresponds to the theoretical ground state of the given system.
The value of the stress after geometry optimization gives an indication of the magnitude of the
elastic effects associated with the supercell approximation.
Mixture atoms
Atomic sites in a disordered crystal can be described in terms of a hybrid atom, which consists of two or
more element types. The relative concentrations can be set for any number of atoms, where the total
concentration must be 100%. The mixture atoms description is the most often used representation of
solid solutions, metallic alloys, disordered minerals, and so on.
Technical limitations
Current implementation of the molecular dynamics (MD) algorithms in CASTEP is not yet compatible
with the virtual crystal approximation (VCA) formalism, which is used to model disorder in solids. None
of the MD schemes (NVE, NPT, and so on) work for disordered crystals in the current version.
The following properties are unavailable for disordered systems:
Population analysis
Vibrational properties (phonon dispersion, phonon density of states)
Optical properties
In addition, geometry optimization based on damped MD is not available (since MD itself is not
allowed).
Atom occupancy cannot be used with the current CASTEP implementation of the virtual crystal
approximation (VCA). Partial atom occupancy can be considered as a special case of a mixture atom,
where one part of the mixture is an atom of "nothing". This case is not supported for a number of
technical and fundamental reasons.
The limitations listed above fall into three categories:
Properties that are ill-defined or do not make sense at all for mixture atoms (population analysis,
vibrational properties).
Features where the calculations might be possible, but which require more theoretical work (optics,
partial atomic occupancies).
Features that are blocked for purely technical reasons (molecular dynamics).
These limitations are imposed by the user interface; if you attempt a run using a combination of settings
which is not allowed, a message box explaining why the run cannot be carried out will be displayed when
you select Run or Save Files.
Note: CASTEP does not support all possible permutations of electronic settings for all possible
properties requests. In cases where the input settings are contradictory the Materials Studio client
will display messages explaining the problem. However, sometimes the combinations selected during
the setup of a CASTEP calculation are impossible to satisfy. The main source of such contradictions is
in the type of pseudopotentials used. For example, NMR calculations and core level spectroscopy
calculations require the use of on-the-fly generated (OTFG) pseudopotentials for at least some
elements. On the other hand, nonlocal exchange-correlation functionals (for example B3LYP) or linear
response phonon calculations are limited to norm-conserving pseudopotentials. It is therefore
impossible to use the same settings for requesting, for example, phonon dispersion using linear
response and core level spectra, or to calculate band structures with B3LYP and request NMR
chemical shifts.
Tip: Accurate calculations for 3D periodic systems may require a large number of k-points if the lattice
has a strongly skewed unit cell (for example, some cell angles are very small). In such cases it might be
beneficial for the speed of calculations to convert the structure to reduced (Niggli) cell representation
using the Redefine Lattice dialog.
magnetic metallic systems usually requires more accurate calculation settings than those defined by the
Express quality setting.
Note: The Express setting cannot be translated into the Coarse - Ultra-fine accuracy scale.
To help you select appropriate values for CASTEP parameters, four predefined parameter sets are
available. The predefined sets vary in quality from Coarse, suitable for quick approximate calculations, to
Ultra-fine, for calculations where high precision is required.
To select a quality
1. Choose Modules | CASTEP | Calculation from the menu bar to display the CASTEP Calculation
dialog.
2. Select the Setup tab.
3. Choose a Quality setting from the dropdown list to select one of the four predefined parameter
sets.
When you choose a Quality setting, the parameters are set according to the task and the system.
The values of the Energy cutoff that correspond to the Coarse, Medium, and Fine settings are element
specific and are taken from the pseudopotential files stored in the database. These values were
determined from convergence tests for single atoms and diatomic molecules and they roughly
correspond to total energy convergences of 2, 0.3, and 0.1 eV atom-1, respectively. The highest value for
the selected quality setting among the elements present in the 3D structure document is accepted.
The Quality setting also affects the values of the convergence tolerances for tasks such as geometry
optimization, elastic constants calculations, and transition-state searching.
Note: The values of force, stress, and displacement tolerances for the Geometry Optimization and
Elastic Constants tasks are set to 100 for ExpressQuality. This preset value instructs CASTEP to ignore
these tolerances and consider only the energy tolerance as a convergence criterion.
Geometry optimization
Quality Express Coarse Medium Fine Ultra-fine
Energy tolerance (eV/atom) 5.0e-5 2.0e-5 1.0e-5 5.0e-6
Energy tolerance (eV/cell) 0.001
Max. force tolerance (eV/Å) 100 0.1 0.05 0.03 0.01
Max. stress tolerance (GPa) 100 0.2 0.1 0.05 0.02
Max. displacement tolerance (Å) 100 5.0e-3 2.0e-3 1.0e-3 5.0e-4
BFGS Use line search No Yes Yes Yes Yes
Note: The values of force, stress, and displacement tolerances are set to 100 for ExpressQuality. This
preset value instructs CASTEP to ignore these tolerances and consider only the energy tolerance as a
convergence criterion.
The individual tolerances may need to be adjusted in some cases:
The Max. force tolerance can be reduced for molecular systems (isolated molecules, molecules on
surfaces, molecular crystals, and so on)
The Max. stress tolerance is generally expressed as a fraction of the bulk modulus of the material, so
it should be smaller for soft materials. However, in practice, cell optimization of soft substances (for
example, molecular crystals) is difficult and might not be stable enough to impose very stringent
convergence criteria.
Elastic constants
Quality Express Coarse Medium Fine Ultra-fine
Energy tolerance (eV/atom) 1.0e-5 4.0e-6 2.0e-6 1.0e-6
Energy tolerance (eV/cell) 0.001
Max. force tolerance (eV/Å) 100 0.02 0.01 0.006 0.002
Max. displacement tolerance (Å) 100 1.0e-3 4.0e-4 2.0e-4 1.0e-4
Note: The values of force, and displacement tolerances are set to 100 for ExpressQuality. This preset
value instructs CASTEP to ignore these tolerances and consider only the energy tolerance as a
convergence criterion.
Note: These tolerances are five times more strict than the ones used for the Geometry Optimization
task.
TS Search
Quality Express Coarse Medium Fine Ultra-fine
RMS convergence (eV/Å) 0.1 0.5 0.25 0.05 0.01
Electric fields
An external electric field can be applied to the system. This causes a force to act on each atom which is
proportional to the charge on that atom. This can be used to calculate the polarizability of a dielectric
material.
To apply an electric field
1. Choose Modules | CASTEP | Calculation from the menu bar to display the CASTEP Calculation
dialog.
2. Select the Electronic tab and click the More... button to access the CASTEP Electronic Options
dialog.
3. Select the Potentials tab.
4. Specify the strength of the Electric field to apply.
5. Enter the Field direction in lattice vectors (A B C).
Setting the exchange-correlation functional
CASTEP supports one local exchange-correlation functional, LDA, three gradient-corrected exchange-
correlation functionals, GGA, and a set of nonlocal functionals for the self-consistent total energy
calculations.
LDA functional
CA-PZ is the only local (LDA) functional available in CASTEP. It is based on the Ceperley and Alder (1980)
data as parameterized by Perdew and Zunger (1981).
GGA functionals
GGA functionals provide a better overall description of the electronic subsystem than the LDA
functionals. The LDA description tends to overbind atoms, causing underestimation of the bond lengths
and the cell volume by a few percent and overestimation of the bulk modulus. GGA corrects this error
but may underbind instead, leading to slightly long bond lengths.
PBE (Perdew et al., 1996) is the default exchange-correlation functional. PBE is especially suitable for
studies of molecules interacting with metal surfaces, although it is also fairly reliable for bulk
calculations. Use the PW91 (Perdew et al., 1992) functional for comparisons with literature data, as it is
the most widely used GGA functional. The results obtained with the PW91 functional are usually very
similar to those obtained with PBE. Another nonlocal functional, RPBE (Hammer et al., 1999), is a revised
form of the PBE functional designed to improve the description of metallic surfaces. One of the latest
GGA potentials, WC, was developed by Wu and Cohen (2006). The authors claim significant
improvements for lattice constants, crystal structures, and metal surface energies over the most
popular Perdew-Burke-Ernzerhof (PBE) GGA. Another recent functional, PBEsol, was developed
specifically to improve the description of exchange in solids, resulting in better structures and energetics
for densely packed solids and their surfaces (Perdew et al., 2008).
Meta-GGA functional
In addition to the generalized gradient (GGA) functionals, which depend on the local density and its
gradient, CASTEP can handle functionals that depend on the kinetic energy density. The current
supported functional is RSCAN (Bartok and Yates, 2019), which is an improved regularized version of the
SCAN ("Strongly Constrained and Appropriately Normed") functional developed by Sun et al., 2015.
Meta-GGA functionals are considered to be more accurate than pure GGA, and the cost of such
calculations is significantly lower than for nonlocal functionals.
Note: There are limitations on the use of meta-GGA in CASTEP. The meta-GGA functional is not
compatible with spin-orbit coupling, J-coupling, linear-response phonons, or polarizability
calculations. You cannot use the formalism with mixture atoms. You can only use on-the-fly-
generated pseudopotentials.
The preferred functionals are sX-LDA, PBE0, B3LYP, and HSE06. The other functionals allow testing
against other packages that implement Hartree-Fock schemes for solids.
Note: There are limitations on the use of this approach: you can only use nonlocal exchange for
insulators using the all-bands minimizer (not density mixing). It is not compatible with NMR, phonon,
or polarizability calculations. You cannot use the formalism with mixture atoms or ultrasoft
potentials.
To select the exchange-correlation functional
1. Choose Modules | CASTEP | Calculation from the menu bar.
2. Select the Setup tab.
3. Choose a local (LDA), a gradient corrected (GGA) or nonlocal functional from the first dropdown list
for the Functional.
4. Select the specific functional from the second dropdown list.
Setting Up Pseudopotentials
Ultrasoft and Norm-Conserving Pseudopotentials
CASTEP uses on-the-fly generated ultrasoft pseudopotentials (USP) by default. These are generally more
accurate and more efficient than norm-conserving pseudopotentials (NCP).
USP implementation allows you to run CASTEP calculations with a lower energy cutoff than their NCP
counterparts producing a clear advantage in terms of the calculation time. However, USP formalism is
more complex and becomes close to intractable for such complex concepts as linear response
implementation for phonons or NMR properties, or for nonlocal exchange-correlation functionals. There
is a number of tasks and properties that CASTEP can address only with norm-conserving potentials.
There are two types of pseudopotentials available in Materials Studio: tabulated and generated on-the-
fly.
Pseudopotentials Generated On-the-Fly
CASTEP server code allows you to generate pseudopotentials on the fly, that is, you can provide the
parameters that govern the generation rather than a tabulated file from the database. This approach
has a number of advantages; for example:
It uses the same exchange-correlation functional in the atomic and solid-state calculations.
It is possible to generate "softer" or "harder" potentials by changing the core radius.
It is possible to study excited configurations with a core hole, and so on.
The latest set of OTFG settings for ultrasoft pseudopotentials was developed to minimize the error with
respect to fully converged all-electron DFT calculations. The error achieved by this set is 0.4 meV/atom,
which puts CASTEP among the most accurate pseudopotential codes available. A full definition of the
test framework and the meaning of the error is given by Lejaeghere et al. (2014), and on the website of
the Delta project: http://molmod.ugent.be/deltacodesdft. The files that correspond to this set have the
_2017R2 suffix. This is the recommended set of OTFG ultrasoft potentials. Use files with the _
2017R2ncp suffix when you require norm-conserving potentials. This set has a 1.1 meV/atom error in
the Delta project tests. In addition, provides a set of "high throughput" ultrasoft OTFG settings for
scripting. This QC5 set generally requires lower cutoff energy and offers faster but less accurate
calculations (1.9 meV/atom error in the Delta project tests).
Note: It is not recommended to use the PW91 exchange-correlation functional when requesting on-
the-fly generation of pseudopotentials. PBE, RPBE, WC, BLYP, or PBESOL are better options when you
require a GGA functional.
This scheme was validated for all elements of the Periodic Table and the files themselves include the
results of the tests. You can also create customized versions of the pseudopotential generation settings.
In which case, test the resulting potentials thoroughly for convergence and transferability by performing
calculations for a variety of systems and at different energy cutoffs. For details of the format of the on-
the-fly generation settings file, see CASTEP file formats - OTFG.
On-the-fly generation of pseudopotentials involves the solution of the electronic structure of an atom
and a pseudoatom. You can perform this step either in a non-relativistic (Schroedinger) framework,
using one of the available scalar relativistic approaches (Koelling-Harmon or ZORA), or using a fully
relativistic atomic solver (Dirac). It is important to use a scalar relativistic treatment for improved
accuracy of calculations for heavy elements.
Notes:
You must use the Dirac solver when you request spin-orbit coupling.
You can only compute J-coupling using the Schroedinger or ZORA approach.
You can only perform CASTEP NMR calculations using the on-the-fly generation scheme, since the
available potentials in the database do not contain sufficient information about all-electron
projectors.
Tabulated Pseudopotentials
Tabulated USP files in the database have a .usp or .uspcc extension, tabulated norm-conserving files
have a .recpot extension.
Note: Use different tabulated ultrasoft potentials for LDA and GGA exchange-correlation functionals.
The naming convention in the database is <element name>_<ID>.usp[cc] for LDA
functionals and <element name>_<ID>PBE.usp[cc] for GGA functionals. By default,
Materials Studio tries to find the best fit of the potentials to the exchange-correlation potentials,
although you can override this behavior.
The database of NCPs available in Materials Studio is rather old and in some cases might contain
potentials that have not been tested sufficiently. An alternative is to use on-the-fly generated
pseudopotentials that are more modern and more extensively tested.
Besides changing the global selection of pseudopotentials, it is possible to change the individual
selections for each element according to the selected value of the exchange-correlation functional.
Tip: Individual selections are also useful for elements such as oxygen and silicon where additional soft
versions of the USPs are available. These require a smaller energy cutoff but may be less transferable
than the standard potentials.
Therefore, CASTEP derives a finer FFT grid in both of these cases, compared with other exchange-
correlation functionals or norm-conserving potentials.
The table below shows the factor by which the actual grid differs from the ideal one depending on (i)
Derived grid setting and (ii) the use of either the PW91 functional or ultrasoft pseudopotentials.
Tip: Use the Fine setting (or better) for high-accuracy calculations, especially when the cell
optimization is required.
Note: Explicit selection of Γ-point sampling is not recommended. CASTEP will automatically use this
option if the cell is very large in real space, so that the Brillouin zone is small and no sampling is
required.
In all other cases, the Γ-point is the least representative of all Brillouin zone points and using it as the
only sampling point can distort the results severely.
Puska (2000) showed, for example, that there is a 1-2 eV error in the vacancy formation energy in
silicon when the Γ-point is used for cells containing up to 128 atoms.
Finer control over k-point sets can be exercised by using the k-points tab on the CASTEP Electronic
Options dialog. This tab allows you to enter the actual Monkhorst-Pack mesh parameters by selecting
the Custom grid parameters option.
CASTEP automatically optimizes the mesh parameters according to the point group symmetry of the 3D
model. For example, in a cubic crystal the even mesh parameters, N, will generate the set with as many
k-points as the one with the odd mesh parameters, N-1. Since the former set provides better sampling,
it will be selected automatically when either Quality or k-point separation is used to define the k-point
set. Custom grid parameters is the only option which allows you to specify a suboptimal mesh with odd
divisions.
The quality of the k-point sampling is particularly important for metallic systems, where rapid changes in
electronic structure may occur along the energy band that crosses the Fermi level. Insulators or
semiconductors, even when they are treated using variable occupation numbers for electronic states,
are less sensitive to the quality of k-point sampling. The default settings used by CASTEP are designed to
give accurate sampling for metallic systems. This means that you can get good results for insulators and
semiconductors with a slightly less fine k-point mesh than the default.
Note: The total energy is not guaranteed to decrease as more k-points are added. Therefore, when
carrying out convergence testing, you should strive to find a set of k-points such that further
improvements do not alter the total energy beyond the tolerance level you require. However, the
energy is not likely to converge smoothly and oscillations are to be expected.
scheme is especially good for metallic systems, where speedups for metal surfaces compared to
conjugate gradient schemes are in the region of 10-20.
The only case where density mixing may not improve performance is for molecule in a box calculations.
The default density mixing settings use Pulay mixing and conjugate-gradient minimization of each
electronic state. You should only attempt to change these parameters if SCF convergence is very poor.
Sometimes it helps to reduce the length of the DIIS history from the default value of 20 to a smaller
value (5-7). It might also be helpful to decrease the mixing amplitude from the default value of 0.5 to 0.1-
0.2.
Variable electronic states occupancies
By default CASTEP uses variable electronic occupancies, thus effectively treating all systems as metallic.
This is recommended, as it speeds up density mixing optimization, even for systems with large band
gaps. The number of empty bands should be sufficiently large to cater for nearly degenerate bands close
to the Fermi level. This is relevant for transition or rare earth metal compounds, where narrow d or f
bands can be pinned at the Fermi level.
If the number of bands used for such a system is insufficient, the SCF convergence will be very slow and
probably oscillatory. Occupation numbers of the highest electronic states as reported in the .castep
file are likely to be noticeably nonzero for at least some k-points.
SCF convergence with the Density mixing minimization scheme can sometimes be poor for metallic
systems. If this is the case, the alternative All Bands/EDFT scheme, which is based on the ensemble
density-functional theory (Marzari et al., 1997) offers a more robust alternative.
Tip: Slow SCF convergence is often indicative of an insufficient number of empty bands, especially in
spin-polarized calculations. To check if this is the cause, inspect the occupancies of the highest energy
electronic states. They should be very close to zero for all k-points in a calculation which is setup
correctly.
Dipole corrections
By default CASTEP does not perform dipole corrections, however, for slab and single molecule systems
with P1 symmetry either a Self-consistent or Non self-consistent dipole correction can be applied. The
self-consistent scheme is based on the method suggested by Neugebauer and Scheffler (1992) where the
correction potential is recalculated at each SCF step. The non self-consistent scheme is based on the
approach of Yeh and Berkowitz (1999) where a dipole correction is added after the SCF has converged. In
this scenario only the total energy and its gradients are corrected, not the electrostatic potential. As a
result, this type of correction is not suitable for electrostatic potential studies such as workfunction
calculations.
Dipole corrections can be essential in eliminating nonphysical electrostatic interactions between periodic
images, improving accuracy of calculated adsorption energies for molecules on surfaces, for example.
Tip: It is recommended to use the All Bands/EDFT electronic minimization scheme when applying a
self-consistent dipole correction to an elongated cell of a slab representing a metal surface; the
Density mixing minimization scheme may fail to converge for such systems.
For a detailed description of the quality settings please refer to the Setting the quality of a calculation
topic.
Geometry optimization algorithm
The BFGS algorithm is the recommended algorithm for minimizing structures. The damped molecular
dynamics approach is always slower and does not allow cell relaxation. The only case when damped
dynamics may be appropriate is in systems with a very flat potential energy surface. BFGS can be
extremely slow in such cases and can get stuck in a local minimum.
The TPSD geometry optimization algorithm is a viable alternative for systems where the potential energy
surface deviates significantly from the quadratic form. TPSD offers robust convergence and is
particularly recommended when user-supplied constraints on lattice parameters are applied during cell
optimization (for example, when optimizing a solid-solid interface in a direction perpendicular to the
interface plane).
Constraints
Geometry optimization in CASTEP obeys constraints on fractional atomic positions imposed using the
Materials Visualizer. You can set such constraints using the Edit Constraints dialog, accessible from the
Modify menu. Note that atoms must be completely fixed; CASTEP does not support partial constraints
on the x, y, and z components of Cartesian atom positions.
Tip: At present, you must check the Fix fractional position checkbox on the Atom tab of the Edit
Constraints dialog in order to apply fixed atom constraints in CASTEP. This has the effect of fixing both
fractional and Cartesian positions in a fixed cell calculation, and fixing fractional positions in a variable
cell run. In the current version of Materials Studio, checking the Fix Cartesian position checkbox has
no effect on CASTEP calculations.
More general linear constraints in CASTEP can be specified manually by editing the .cell file.
Note: The default setting in CASTEP is to optimize atomic coordinates only. You must choose an
option from the Cell optimization dropdown list on the Minimizer tab of the Geometry Optimization
dialog if cell optimization is required.
Non-linear constraints
Geometry optimization using delocalized internals in CASTEP obeys constraints on interatomic
distances, bonds, and/or torsions imposed using the Materials Visualizer. You can set such constraints
using the Edit Constraints dialog, accessible from the Modify menu.
Notes:
All non-linear constraints are only taken into account if you choose the delocalized internal
minimizer, by checking the Use delocalized internals checkbox on the Options tab of the CASTEP
Geometry Optimization dialog.
The Damped MD algorithm for geometry optimization acknowledges only constrained bonds
(distances) between atoms, all other constraints (torsions and angles) are ignored.
Non-linear constraints cannot be applied in conjunction with symmetry constraints. Thus, if your
crystal has symmetry other than P1, you will be prompted to convert the crystal symmetry to P1
before you can continue.
Non-linear constraints cannot currently be applied to disordered systems.
Potential pitfalls
One of the most difficult parts of geometry optimization in CASTEP is cell optimization. It is important to
ensure that you use a finite basis set correction in such calculations, and preferably a fine quality level.
The accuracy of the stresses which are used in the cell optimization algorithm are crucial for the success
of geometry optimization. Certain systems, for example, rare-earth compounds with highly localized
electronic states, require very stringent criteria for SCF convergence in order to converge the stress
tensor. This is particularly true for the Density mixing minimizer.
If cell optimization fails with the following message in the output file:
"BFGS: Geometry optimization failed to converge after ... steps",
Choosing an ensemble
You can control temperature of a CASTEP dynamics calculation, in order to simulate a system that
exchanges heat with the environment. Under these conditions, the total energy of the system is no
longer conserved and extended forms of molecular dynamics are required.
You can simulate several thermodynamics ensembles with a constant number of particles.
Ensemble Conditions Description
NVE Constant- The Newtonian equations of motion, which conserve the total
volume/constant- energy, are used. The unit cell is fixed.
energy dynamics The temperature is allowed to vary. Energy is the constant of
motion - it can fluctuate slightly, but there should be no
systematic drift.
NVT Constant- The unit cell is fixed. A thermostat maintains a constant
volume/constant- temperature.
temperature The dynamics are modified to allow the system to exchange heat
dynamics with the environment at a controlled temperature. Several
methods of scaling or controlling the temperature are available.
NPH Constant- The size of the unit cell is allowed to vary with a barostat to
pressure/constant maintain constant pressure. The enthalpy is a constant of the
enthalpy dynamics motion.
The size of the unit cell is allowed to vary, while the temperature is
also allowed to vary. The enthalpy, which is a constant of the
motion, is not allowed to vary beyond a set value between
dynamics steps.
NPT Constant- The size of the unit cell is allowed to vary with a barostat to
pressure/constant- maintain constant pressure. A thermostat maintains a constant
temperature temperature.
dynamics Similar to NPH, except that temperature is controlled. Several
methods of scaling or controlling the temperature are available.
To choose an ensemble
1. Choose Modules | CASTEP | Calculation from the Materials Studio menu bar.
2. Select the Setup tab.
3. Click the More... button to open the Dynamics dialog.
4. Select the Dynamics tab.
5. Select the required ensemble from the Ensemble list.
6. Select the Options tab to enter parameters for the selected ensemble.
Constraints
Molecular dynamics in CASTEP obeys constraints on atomic positions and/or interatomic distances
(bonds) imposed through the Materials Visualizer. You can set such constraints using the Edit
Constraints dialog, accessible from the Modify menu.
To fix atomic positions in CASTEP you must check the Fix fractional position checkbox on the Atom tab
of the Edit Constraints dialog.
Tip: To optimize the reactant and product structures as part of the TS Search run, check the Optimize
reactants and products checkbox on the CASTEP Transition State Search dialog.
Note: Elastic Debye temperature and averaged sound velocity are not reported if the calculated
elastic constants do not produce physically meaningful sound velocities in some directions (for
example, if certain Cij values or their combinations are negative when they should be positive). This
can happen for low quality calculations or for calculations that were not preceded by a geometry
optimization (including optimization of the lattice parameters).
A CASTEP elastic constants calculation can also produce the piezoelectric strain tensor if you request
polarization. The calculation uses the same set of optimized structures, with the applied strain tensor to
differentiate numerically calculated polarization with respect to applied strain. This property is even
more sensitive to the accuracy of the calculation than elastic coefficients. In particular, you can improve
results by using extremely tight SCF convergence criteria and verifying convergence of the results with
respect to k-point sampling of the Brillouin zone.
Setting up strain patterns
CASTEP generates the strain patterns automatically, based on the lattice type of the structure. The
maximum number of patterns required is six for a monoclinic structure, while one pattern is sufficient
for cubic cells. Certain symmetries allow you to modify the patterns so that, whenever possible, the
calculation generates volume-conserving strains.
The quality of the calculated values depends strongly on the amplitude of the applied distortions and
the number of points used for each strain pattern. Use at least four steps for each strain to obtain a
statistically reliable linear fit of the stress-strain relationship. Choose the maximum strain amplitude so
that:
It is sufficiently low that the material stays within the linear elasticity regime.
It is sufficiently high that the distortion generates stresses significantly higher than the error level of
the CASTEP calculation (errors from imperfect SCF convergence, for example).
To set up strain patterns
1. Choose Modules | CASTEP | Calculation from the Materials Studio menu bar.
2. Select the Setup tab.
3. Choose the Elastic Constants task.
4. Click More... to open the CASTEP Elastic Constants dialog.
5. Select the Elastic Constants tab.
6. Specify the Number of steps for each strain and Maximum strain amplitude in the appropriate text
boxes.
Convergence criteria
The Quality setting on the Minimizer tab of the CASTEP Geometry Optimization dialog controls the
accuracy of the calculation.
For a detailed description of the quality settings, see Setting the quality of a calculation.
Note: Sometimes the density mixing scheme may not converge. In such cases, switch to an EDFT
minimizer to improve results and choose to Optimize total spin after 1 SCF step.
Note: You can only apply the Hubbard U operator to the states that have an LCAO projector in the
pseudopotential. If there is no projector for any orbital with nonzero U value, CASTEP returns the
following error message:
Hubbard U is requested for the states which do not have LCAO
projectors. Please check HUBBARD_U block in the cell file.
If you think that the states to which you want to apply Hubbard U are really important, use a different
pseudopotential (on-the-fly generated potentials are likely to have more projectors than other
pseudopotentials).
Note: CASTEP calculations with core holes assume by default that an electron is removed from the
system, so it is treated as having a charge of 1 (or one more than specified in the Charge field on the
Setup tab of the CASTEP Calculation dialog). In some circumstances, for example when studying
impurities, it might be more appropriate to consider the system as charge neutral. In order to achieve
this you should specify the Charge of -1 on the Setup tab, in this way the CASTEP input file will have a
zero charge. It is also possible to edit the seedname.param file to specify a fractional charge. The
strength of the core hole can be adjusted by editing the seedname.cell file and modifying the
pseudopotential definition string (for example, change {1s1} to {1s0.5} to use a core hole with
the charge of 0.5).
The result of the transition state search is an 3D Atomistic Collection Document containing reactant,
product, and transition state structures. You can select the "TS" physical system and use the Vibrational
Analysis tool to verify the nature of the imaginary mode associated with the reaction.
Note: The reaction rate coefficient can be calculated only if the transition state has exactly one
imaginary frequency with its eigenvector along the reaction path.
The 3D Atomistic Collection Document generated will contain all the necessary information to perform
the reaction rate calculation itself.
To set up a reaction kinetics calculation
1. Prepare two documents containing reactant and product structures. Set any necessary fixed atom
constraints (fixed atoms should be present in both documents and should have the same
coordinates).
2. Follow the steps required to set up transition state search calculation.
3. Check the Optimize reactants and products checkbox on the CASTEP Transition State Search dialog,
unless you have optimized both structures prior to constructing reaction preview document.
It is recommended to select Ultra-fine for the convergence tolerance for geometry optimization by
changing the task to Geometry Optimization, opening CASTEP Geometry Optimization dialog,
editing convergence settings, and then changing the task again to TS Search. Such an accurate
setting may be necessary to eliminate spurious imaginary modes in the vibrational analysis.
4. Check the Calculate partition functions checkbox on the CASTEP Transition State Search dialog.
5. It is recommended to impose more strict convergence criteria than the default values; lack of
convergence can result in spurious imaginary modes of the transition state itself.
Densities of states
NMR shielding tensors and electric field gradients
Vibrational properties
Polarizabilities and IR absorption intensities
Bond and band populations
Note: It is recommended that only the Γ-point is used for electronic calculations on molecular
systems. In addition, the all bands/EDFT electronic minimizer is known to be more efficient for studies
of isolated molecules in supercell geometry than the density mixing minimizer. Therefore, it is
recommended that you change the Electronic minimizer setting to All Bands/EDFT on the SCF tab of
the CASTEP Electronic Options dialog when performing such calculations.
Tip: Calculations for "molecule in a box" systems that do not require geometry optimization can be
sped up if the molecular symmetry is utilized. Use the Find Symmetry tool to find and apply the
symmetry of the molecule to the supercell.
To create a molecule in a box
1. Either import a structure from a pre-existing file or construct a new molecule using the sketching
tools in the Materials Visualizer.
2. Construct a crystal using the crystal building tools in the Materials Visualizer.
3. If your system does not require geometry optimization, find and impose the symmetry of the
system using the Find Symmetry tool in the Materials Visualizer.
Tip: The Population analysis checkbox on the Properties tab of the CASTEP Calculation dialog is
checked by default.
CASTEP produces electronic properties data by running separate non-self-consistent calculations. This
means that Band structure, Density of states, Optical properties, and Vibrational properties
calculations all generate their own output files. Population analysis, Polarization, and Stress properties
are calculated as part of the main run, each storing their output in a .castep file. CASTEP generates
the Electric field response either as part of a Vibrational properties calculation, in which case the output
is in the appropriate .castep file, or as a separate calculation.
Requesting band structure
Calculating band structure properties produces electronic energies along high symmetry directions in
the Brillouin zone. The standard path for each lattice type is taken from Bradley and Cracknell (1972).
The path can be modified using the CASTEP Band Structure Options dialog. The density of points along
the path, which affects the appearance of the resulting chart, can be controlled by the selection for the
k-point set dropdown list or the approximate k-point separation can be manually specified in the
Separation text box, in Å-1, on the Properties tab of the CASTEP Calculation dialog. Conduction band
states can be included by specifying a nonzero value in the Empty bands text box on the same tab.
Note: You are not required to use the same exchange-correlation functional in the band structure
calculation as in the main calculation. This allows you to select a nonlocal screened-exchange
functional to study the electronic structure after performing, for example geometry optimization with
a variable cell using the GGA functional.
Note: CASTEP does not support the use of nonlocal functionals (screened exchange, exact exchange,
and so on) for ultrasoft pseudopotentials, that is, nonlocal functionals can only be requested if all the
pseudopotentials for the calculation are specified as norm-conserving.
To calculate band structure
1. Choose Modules | CASTEP | Calculation from the Materials Studio menu bar.
2. Select the Properties tab.
3. Check the Band structure option on the Properties list.
4. If necessary, modify the Number of empty bands to include more or less states in the conduction
band.
5. Optionally, adjust the density of k-points along the high symmetry directions by using k-point set
dropdown list.
6. Click the More... button to open the CASTEP Band Structure Options dialog.
7. Click the Path... button to open the Brillouin Zone Path dialog. This dialog allows you to inspect and
modify the Brillouin zone path used for the band structure calculation.
8. On the CASTEP Band Structure Options dialog check the Use separate XC functional for band
structure calculation checkbox to use a nonlocal functional for the electronic structure calculation.
9. Select the alternative Functional you wish to use from the dropdown list.
Requesting core level spectroscopy
Calculating core level spectra produces electronic energies on the Monkhorst-Pack mesh of k-points and
the matrix elements for electronic interband transitions. The CASTEP Analysis dialog can be used to
generate observable core level spectra based on this information. Detailed analysis of the impact of
calculation parameters (k-point set, number of empty bands, and so on) on the accuracy of the spectra
generated can be found in the core level spectroscopy theory topic.
To calculate core level spectra
1. Choose Modules | CASTEP | Calculation from the menu bar.
2. Select the Properties tab.
3. Check the Core level spectroscopy option on the properties list.
4. If necessary, modify the number Energy range above the Fermi level for which you wish to calculate
the spectrum.
5. Optionally, adjust the density of Monkhorst-Pack k-points used for the optical properties
calculation by using k-point set dropdown list.
6. Click the More... button to open the CASTEP Core Level Spectroscopy Options dialog.
7. Select the Core level spectroscopy k-points option you want to use and the Band energy tolerance.
Note: CASTEP calculations with core holes assume by default that an electron is removed from the
system, so it is treated as having a charge of 1 (or one more than specified in the Charge field on the
Setup tab of the CASTEP Calculation dialog). In some circumstances, for example when studying
impurities, it might be more appropriate to consider the system as charge neutral. In order to achieve
this you should specify the Charge of -1 on the Setup tab, in this way the CASTEP input file will have a
zero charge. It is also possible to edit the seedname.param file to specify a fractional charge. The
strength of the core hole can be adjusted by editing the seedname.cell file and modifying the
pseudopotential definition string (for example, change {1s1} to {1s0.5} to use a core hole with
the charge of 0.5).
Note: CASTEP does not support the use of nonlocal functionals (screened exchange, exact exchange,
and so on) for ultrasoft pseudopotentials, that is, nonlocal functionals can only be requested if all the
pseudopotentials for the calculation are specified as norm-conserving.
Note: To obtain representative density of states it is recommended to use a k-point set which is
either the same or finer quality than the one used in the SCF calculations.
Tip: When only one set with the correct name is present in the structure, Materials Studio assumes
that the rest of the atoms should be treated as a second set. Thus, in the case of a molecule-surface
interaction, it is sufficient to create a single set that comprises all the atoms in the molecule.
The isotropic shielding, σiso, is given by one third of the trace of σ(r). From the symmetric component of
the shielding tensor, the chemical shielding anisotropy, σaniso and the asymmetry parameter, η, can be
defined in terms of the principal components of the shielding tensor:
and:
where:
For values of σaniso close to zero the asymmetry parameter will be ill-defined.
Electric field gradient
The interaction of a quadrupolar nuclei with an external magnetic field can be characterized by the
quadrupolar coupling constant, CQ, and the asymmetry parameter, ηQ. If the principal components of
the traceless electric field gradient tensor are labeled, Vxx, Vyy, and Vzz such that:
then:
For very small values of Vzz the asymmetry parameter will be ill-defined.
The value of the quadrupolar moment depends on the actual isotope; CASTEP makes a suggestion
based on the most typical isotope used in solid state NMR experiments. The value of Q affects only the
quadrupolar coupling constant, CQ, calculated and printed out by CASTEP. If you find that the wrong
value of Q was used you can rescale the generatedCQ constant using a different Q value.
Note: The list of default values is based on the compilation by Pyykkö (2008). CASTEP versions prior
to the 5.0 release used an older compilation by Harris (1996), so some results for quadrupolar
coupling constants may have changed in later versions.
J-couplings
NMR J-coupling or nuclear spin-spin coupling is an indirect interaction of the nuclear magnetic moments
mediated by the bonding electrons. It is manifested as the fine structure in NMR spectra, providing a
direct measure of bond strength and a map of the connectivities of the system.
Induced magnetization density and current density are expected to be short-ranged. This forms the
basis of using J-coupling as a tool for probing the strength of interatomic bonds. However, in a periodic
calculation the perturbing nucleus can be viewed as similar to a defect in a defect calculation. Hence,
sometimes it might be necessary to construct a supercell which is large enough to inhibit the interaction
between the periodic defects or perturbations. Convergence with respect to the cell size should
represent an important test in establishing accuracy of J-coupling calculations in either periodic systems
or in molecule in a box calculations.
Note: J-coupling can be calculated only for non-magnetic systems. J-coupling calculations are
supported only for systems with P1 symmetry. If your crystal has symmetry other than P1, you will
be prompted to convert the crystal symmetry to P1 before you can continue.
Tip: It is not recommended to use PW91 exchange-correlation functional when requesting NMR
properties. PW91 functional may result in sharp real-space features of on-the-fly-generated
pseudopotentials; PBE, RPBE, WC, BLYP, or PBESOL are better options when a GGA functional is
required.
EPR G-tensor
Electron Spin Resonance (EPR - also known as Electron Paramagnetic Resonance) is a powerful
experimental probe of the structure of defects in solids. EPR spectra of spin ½ centers are made up of
two contributions:
the hyperfine parameters
the g-tensor
The g-tensor arises from the interaction of the electronic spin with external magnetic field. This term
plays a somewhat similar role to the shielding in NMR; induced electronic currents in the sample modify
the g tensor from its vacuum value.
EPR spectrum can be modeled using the following effective Hamiltonian, bilinear in the total electron
spin S, and the applied uniform magnetic field or nuclear spins, B and II, respectively:
where:
atomic units are used
Tip: Select Fix fractional position on the Atom tab of the Edit Constraints dialog, accessible from the
Modify menu.
frequencies. An advantage of the finite displacements method is that it is also possible to investigate
phonon anharmonicity by increasing the amplitude of the displacements. A further important
advantage is that you can use efficient ultrasoft potentials to obtain vibrational frequencies.
The formalism is based on evaluation of the dynamical matrix at a (coarse) commensurate grid of q-
points at the cost of creating a supercell. Then the dynamical matrix is re-interpolated onto a requested
group of q-points. There are two available approaches to dynamical matrix calculation:
Create a large supercell with the size defined by a real space cutoff parameter; the supercell size then
dictates a coarse grid.
Create a set of non-diagonal supercells (Lloyd-Williams and Monserrat, 2015) which are
commensurate with a q-point group of a coarse grid. In this case, the grid spacing for interpolation
defines the coarse grid.
Limitations of the Finite Displacements Implementation in CASTEP
There is a limitation with the implementation of the finite displacements method in CASTEP related to
the treatment of the non-analytical term responsible for the LO-TO splitting in insulators. This term is
calculated using linear response with an electric field as a perturbation. Therefore, the computational
limitation applies to finite displacements calculations if you request LO-TO splitting.
Note: An LO-TO calculation imposes the restriction that applies when you use the linear response
technique. You might want to proceed with finite displacement calculations without LO-TO splitting if
this restriction affects your study. This case only reliably produces TO modes when there is LO-TO
degeneracy in the results.
9. If you want to inspect or modify the Brillouin zone path used for the phonon dispersion calculation,
click Path... to display the Brillouin Zone Path dialog and alter the path segments accordingly.
10. Optionally, use the partial Hessian approach to accelerate the calculation. Apply fixed atom
constraints to the atoms that are not essential to the calculation; for example, the atoms in a
molecule on surface calculations.
Requesting Phonon Density of States
Calculating the phonon density of states produces phonon frequencies and eigenvectors on the regular
Monkhorst-Pack mesh of q-vectors. CASTEP requires this information to produce either the total or
projected (partial) phonon density of states and, thus, to calculate thermodynamic properties. The
Quality setting in the Density of states section on the CASTEP Phonon Properties Options dialog
controls the quality of the q-vector.
Note: When you request phonon density of states CASTEP also calculates atomic displacement
parameters, allowing assignment of temperature factors during analysis.
To calculate the density of states
1. Choose Modules | CASTEP | Calculation from the menu bar to display the CASTEP Calculation
dialog.
2. Select the Properties tab.
3. Select Phonons in the list of properties, then select Density of states below the list.
4. If required, select Calculate LO-TO splitting.
5. Select the technique to use to calculate the phonon frequencies as the Method.
6. Click More... to display the CASTEP Phonon Properties Options dialog.
7. For linear response calculations, specify the convergence criterion for the electronic eigenvalues in
Convergence tolerance.
If you do not want to carry out a direct calculation, select Use interpolation and specify the q-point
separation in q-vector grid spacing for interpolation.
For finite displacement calculations, specify the cutoff radius to use to construct the appropriate
supercell in Supercell defined by cutoff radius.
Optionally, use the Quality control in the Density of states section to adjust the density of
Monkhorst-Pack q-vectors used for the phonon density of states calculation.
8. If you want to inspect or modify the setup of the Monkhorst-Pack grid, click More... to display the
Phonon Density of States Options dialog and alter the settings accordingly.
9. Optionally, use the partial Hessian approach accelerate the calculation. Apply fixed atom constraints
to the atoms that are not essential to the calculation; for example, the substrate atoms in a
molecule on surface calculations.
Tip: If you choose quality gamma frequencies, then CASTEP only calculates eigenvectors at the Γ-
point, which determines most of the experimental data on IR or Raman spectra. This can accelerate
finite displacement calculations tremendously as no supercell is constructed for this case. However,
the accuracy of thermodynamic properties obtained with this single q-vector is very poor as it does
not take into account any acoustic degrees of freedom.
bands/EDFT electronic minimizer is known to be more efficient for studies of isolated molecules in
supercell geometry than the density mixing minimizer. Therefore, change the Electronic minimizer to All
Bands/EDFT on the SCF tab of the CASTEP Electronic Options dialog when performing such calculations.
Tip: Calculations for "molecule in a box" systems that do not require geometry optimization can be
sped up if the molecular symmetry is utilized. Use the Find Symmetry tool to find and apply the
symmetry of the molecule to the supercell.
Note: Do not calculate LO-TO splitting for a "molecule in a box" system. There is no physical meaning
to this term when studying molecular, rather than crystalline, vibrations.
The ultimate convergence test would require geometry optimization performed for a denser k-point
grid, followed by a new phonon calculation - this is the only way to determine whether the k-point
sampling has an effect on the accuracy of the calculated vibrational frequencies.
Note: Phonon spectra in metallic compounds are very sensitive to the details of the Fermi surface. To
obtain reasonable spectra using the linear response method for metals, the k-point sampling must be
extremely dense: much denser than the Fine mesh. Consider specifying a value of less than 0.02 for
the Separation parameter on the k-points tab of the CASTEP Electronic Options dialog. A clear sign of
insufficient quality of k-point sampling is the presence of imaginary acoustic modes near the Γ-point.
Note: It is important to select the correct description of the system as the System type when
calculating the electric field response for a molecular system. The wrong system description can
significantly affect the molecular polarizability values obtained.
The piezoelectric strain tensor is related to the piezoelectric stress tensor, through the elastic coefficients
tensor that connects strain and stress responses:
Note: Polarization calculations are not allowed if you use any of the following settings:
Ultra-soft pseudopotentials
Spin-polarized calculation
DFT+U
Spin-orbit coupling with non-collinear spins
External electric field
Mixture atoms
Note: It is not possible to specify the number of atomic orbitals to be used in the population analysis
explicitly in the CASTEP interface. If the spilling parameter reported in the .castep output file is too
large (more than 5%), it may be that the default CASTEP parameters were insufficient for this
calculation. In such circumstances you should edit the .cell file and change the SPECIES_LCAO_
STATES parameters for atomic species, then run the input files again. The SPECIES_LCAO_
STATES parameter for a species should be set such that it results in a complete set of angular
momentum states. For example, to use a single s and single p state SPECIES_LCAO_STATES
must be 2.
To perform population analysis
1. Choose Modules | CASTEP | Calculation from the Materials Studio menu bar.
2. Select the Properties tab.
3. Ensure that the Population analysis option on the properties list is checked.
4. Optionally, modify the Distance cutoff for bond populations field to include more or less atomic
pairs in the calculation.
5. If you wish to generate partial and local densities of states for the valence band, check the Calculate
band populations option.
Requesting solvation energy
Calculating solvation energy in CASTEP involves performing two separate calculations for an isolated
molecule in a supercell: one to simulate the molecule in a vacuum, and another to include the implicit
solvation model. The seedname_Solv.castep file reports the solvation energy.
Note: There are certain limitations for solvation energy calculations:
The system must represent a molecule in a box.
Symmetry of the system must be P1.
Calculations must not include spin-polarization.
Stress calculations are not available.
Note: Use at least a Fine FFT grid Density for solvation calculations; the Precise density is even better.
You can specify the FFT grid Density on the Basis tab of the CASTEP Electronic Options dialog. Also use
an increased value of Augmentation density scaling factor, for example 1.5-2.0 instead of the default
value of 1.0.
An increased FFT density and energy cutoff can help to eliminate convergence errors for the multigrid
solver for electrostatics in a solvent.
To perform a solvation energy calculation
1. Choose Modules | CASTEP | Calculation from the menu bar to display the CASTEP Calculation
dialog.
2. Select the Properties tab.
3. Select the Solvation energy checkbox in the list of properties.
4. Select the Electronic tab on the CASTEP Calculation dialog. Click More...: to open the CASTEP
Electronic Options dialog.
5. Select the Solvent tab.
6. From the Solvent list, select the solvent to use.
7. If you want, you can specify the dielectric constant of the solvent in the Dielectric constant field.
8. Keep selection of the Include apolar term and Account for solute-solvent dispersion repulsion
checkboxes as a robust default setting.
9. If you want, you can specify the surface tension of the solvent in the Solvent surface tension field (in
units of N/m).
Requesting stress analysis
Calculating stress involves performing a stress tensor calculation, even when the cell parameters are
fixed. The stress is a useful measure of the accuracy of the structural description: for example, after
performing geometry optimization with a fixed cell, you can get an idea of how accurate the cell
parameters are from the magnitude of the remaining stress. Similarly, when studying an isolated
molecule in a supercell, the value of the stress provides a guide as to the size of the cell required to
remove unphysical interactions between periodic images.
Note: It is important to use finite basis set correction when calculating the stress tensor (see Setting
up the basis set for details). Choosing either Smart or Always as the Apply finite basis set correction
setting on the Basis tab of the CASTEP Electronic Options dialog will achieve this.
To perform a stress analysis
1. Choose Modules | CASTEP | Calculation from the menu bar to display the CASTEP Calculation
dialog.
2. Select the Properties tab.
3. Check the Stress checkbox in the list of properties.
4. Select the Electronic tab.
5. Click the More... button to display the CASTEP Electronic Options dialog.
6. Select the Basis tab.
7. Select either Smart or Always from the Apply finite basis set correction dropdown list.
Manipulating files
CASTEP is a file-based application which means that all input information is provided in various ASCII text
files and the output is delivered in a mixture of text and binary files. This section describes some file
handling issues which may arise, especially when the CASTEP server is run in a standalone mode and not
through a gateway.
Input files
CASTEP allows you to save input files, for subsequent manual editing instead of running the job directly.
To save input files
1. Choose Modules | CASTEP | Calculation from the Materials Studio menu bar.
2. Click the Files... button to open CASTEP Job Files dialog.
3. Click the Save Files button.
Note: The Save Files button is enabled only if a suitable 3D model document is active.
Only one input file, .param, is displayed in the Project Explorer. This file contains the parameters
specified using the CASTEP interface. However, it also contains a few parameters that allow you to access
functionality which is not supported though the interface. The other input files are hidden, since they
are unlikely to be edited manually.
Input files can be run on a server after they have been edited:
To run CASTEP using an existing set of input files
1. Choose Modules | CASTEP | Calculation from the Materials Studio menu bar.
2. Click the Files... button to open CASTEP Job Files dialog.
3. Click the Run Files button.
Note: The Run Files button is enabled only if a suitable .param file is active.
If your server does not support the gateway protocol, you may have to run CASTEP in standalone mode
using the RunCASTEP.sh or RunCASTEP.bat files provided with the installation.
In such circumstances it is necessary to copy all the input files from the Materials Studio Project folder to
the appropriate directory on the server machine. The scripts required for standalone execution are
provided with accompanying RunCASTEP.Readme files that explain, in detail, how to use them.
Alternatively, you can obtain help by typing one of the following:
RunCASTEP.sh -h
RunCASTEP.sh
Note: The Materials Studio Project folder does not contain the pseudopotential files required to run a
job. When transferring input files to the server, you must therefore determine which pseudopotential
files are required by examining the .cell file. Then you must locate those potential files in Materials
Studio installation, typically in a sub-directory called share\Resources\Quantum\Castep and copy the
required files to the same directory on the server as the rest of the input files.
Output files
Only one output file, .castep, is displayed in the Project Explorer. However, there are many more files
created during a CASTEP run. These are hidden by Materials Studio. All of these output files are placed in
the correct Materials Studio Project folder automatically, when the job is run using the Gateway.
However, if CASTEP is run in standalone mode, the output files must be copied manually from the server
to the appropriate Project folder.
Some output files can be quite large. If these files are not likely to be required to restart the job, then it is
recommended that you delete or archive them periodically.
Restarting a CASTEP calculation
It is not possible to restart a completed CASTEP job automatically using the interface. However, there
are circumstances in which you may wish to do this, for example, if you are running a geometry
optimization which does not converge in the given number of steps, or a you wish to continue your
calculation with manually modified runtime parameters. To restart such jobs you should edit .param
file and add the CONTINUATION keyword to the file.
Note: Any additional changes to this file may make it impossible to restart the CASTEP job
successfully.
Tip: It is possible to remove everything from .param and leave only the CONTINUATION line. This
will ensure that all of the settings remain exactly the same as those from the previous run.
Note: In order to restart a job, the output files from the previous run must be present in the Project
folder. This check and the subsequent transfer of the files to the server happens automatically
whenever CASTEP is used to run an existing file set or to calculate additional properties.
The atoms in the updated structure document have an extra property, ServerAtomName, associated
with them. This shows the atom sequence numbers as used in the CASTEP output files and can be
seen by labeling atoms in the structure with the ServerAtomName property. (This property will be
seen only in structures generated from CASTEP and other similar applications).
All other analysis functions require you to perform certain actions using the CASTEP Analysis dialog.
Note: Only structure updates take place automatically after a CASTEP job is completed and the
results files have been downloaded successfully (this only applies to jobs that are run on a server
through the gateway, not jobs that are run in a standalone mode). All other analysis functions are
performed manually, using the options on the CASTEP Analysis dialog.
CASTEP produces a trajectory (.xtd) file whenever you run a geometry optimization, molecular
dynamics or elastic constants calculation, or a transition state search. In all cases, a history of the
optimization is returned in a file called <name> Trajectory.xtd, where <name> is the name used
for the job description on the Job Control tab of the CASTEP Calculation dialog.
When you run a geometry optimization, two chart documents are generated:
<name> Energies.xcd: Contains a plot of total energy vs. frame for all steps of the calculation
<name> Convergence.xcd: Contains three (possibly four) plots that describe the degree of
convergence of the optimization as a function of frame:
Change in energy vs. frame
Maximum Cartesian force vs. frame
Maximum Cartesian step size vs. frame
Maximum stress tensor component vs. frame (for cell optimization runs)
Because these values span several orders of magnitude, the data are plotted on a log10 scale.
Similar charts are produced for each structure in an elastic constants calculation.
When you run a transition state search, a single chart containing a graph of energy vs. reaction
coordinate for each step of the calculation, <name> TSSearch.xcd, is generated. This includes an
individual graph for each LST or QST 'uphill' cycle as well as a graph for each conjugate gradient
minimization. The CASTEP Transition State Search task topic contains more details.
When you run a molecular dynamics simulation, either two or three chart documents are generated:
<name> Constant.xcd: Contains a plot of the constant of motion vs. simulation time
<name> Temperature.xcd: Contains a plot of temperature vs. simulation time
<name> Pressure.xcd: Contains a plot of pressure vs. simulation time (only produced by the
NPH or NPT ensemble)
Trajectory files can be animated using the controls on the Animation toolbar.
If you have access to the Forcite module, you may find some of its dynamics, statistical, and structural
analysis features useful for analyzing trajectories. Refer to the Using Forcite Analysis topic for details.
Creating a trajectory and chart
When you run CASTEP using the Materials Studio interface, the creation and display of these trajectories
and charts is automatic. If you selected Automatically view output on the CASTEP Job Control Options
dialog, the graphs will be displayed automatically on completion of the run. If you selected Update
graphs on the CASTEP Job Control Options dialog, these graphs will be displayed and updated with
intermediate results throughout the course of the calculation.
To create charts from output files generated by a standalone calculation
1. Download the files from the compute server to your PC using the procedure described in Running
CASTEP in standalone mode. Be sure to retrieve the .check, .castep, and .geom, .md, or
.ts files.
2. Select Modules | CASTEP | Analysis from the menu bar.
3. Select Energy evolution from the list on the CASTEP Analysis dialog.
4. Make sure that the .castep file is the currently active document.
5. Click the View button.
If you used the Materials Studio interface to set up and launch your job, you can regenerate the chart
documents by following a similar procedure.
To regenerate charts
1. Open either the .xsd or .castep file and make it the currently active document.
2. Select Modules | CASTEP | Analysis from the menu bar.
3. Select Energy evolution from the list on the CASTEP Analysis dialog.
4. Click the View button.
Animating the trajectory
You can use the controls on the Animation toolbar to animate the trajectory, stepping through each of
the frames of the document. If the chart is open during the animation, the point corresponding to the
active frame will be highlighted.
Chart Viewer point selection
Alternatively, you can use Chart Viewer point selection to display the trajectory frames of the points you
select from the chart document.
To display a trajectory frame for a point on a chart document
1. Choose the Selection tool on the Chart Viewer toolbar to enter selection mode.
2. Making sure you have the corresponding trajectory document open, select a data point on the
chart document.
3. The corresponding trajectory frame is displayed.
4. Drag the mouse over other data points on the chart document to display their corresponding
trajectory frames.
5. If CASTEP generated multiple chart documents, the related data points will be highlighted on all of
them.
Tip: To make it easier to see the chart and trajectory documents, close all the other documents and
then maximize the sizes of these two documents by selecting Window | Tile Vertically from the menu
bar.
additional separate fields for the density of spin-up (alpha) or spin-down (beta) electrons, as well as a
field for the spin density (the difference between the density of alpha and beta electrons). The latter field
allows you to visualize the spatial distribution of the magnetic moment in a spin-polarized system.
Note: Maxima in the electronic density field correspond to interatomic bonds or to the position of
atoms that have semicore states which are explicitly treated as valence. When there are no semicore
states in the system (this can be verified by analyzing the atomic charges displayed on the Potentials
tab of the CASTEP Electronic Options dialog), the density will display minima at the positions of
atomic nuclei. This is to be expected in the pseudopotential formalism.
To create a density field
1. Choose Modules | CASTEP | Analysis from the menu bar to open the CASTEP Analysis dialog.
2. Select Electron density from the list of properties.
3. Ensure that the 3D Atomistic document to be analyzed is the active document.
4. For spin-polarized calculations, select the type of field required from the Density field dropdown list.
5. Optionally, check the View isosurface on import checkbox to automatically create an isosurface
when the density field is imported.
6. Click the Import button.
You also have the option of creating a plain text file which contains the values of charge (and spin, when
appropriate) density on a regular grid along three lattice vectors. The file has the extension .charg_
frm and is hidden by default. Select Tools | Folder Options... in the Windows File Explorer to open the
Folder Options dialog, then select the View tab and click the Show hidden files and folders radio button
to reveal the hidden file. The format of the file is self-explanatory:
a b c Charge
1 1 1 1.8133362740
2 1 1 1.5071161497
3 1 1 1.1111633102
4 1 1 0.9584556846
Note: The charge and spin density in the .charg_frm file are in the same units as those used by
CASTEP during calculations. The values are normalized to the number of electrons per cell. The
density fields displayed in the 3D model are, however, normalized to the number of electrons per Å3.
To create a plain text density file
1. Choose Modules | CASTEP | Analysis from the menu bar to open the CASTEP Analysis dialog.
2. Select Electron density from the list of properties.
3. Click the Save density to file button.
From - lowest value of the energy of the Nth band at any k-point (in eV)
To - highest value of the energy of the Nth band at any k-point (in eV)
You can artificially change the band gap value for insulators or semiconductors by using the scissors
operator. This parameter effectively describes the difference between the theoretical and experimental
band gap values. When the experimental value is known, you can perform a band structure calculation
to find the theoretical band gap. Then, simply set the scissors operator to the difference between the
two values. The eigenvalues of the conduction band states displayed in the scrollable list will be shifted
upward by the value of the scissors operator.
To create an orbital field
1. Choose Modules | CASTEP | Analysis from the menu bar to open the CASTEP Analysis dialog.
2. Select Orbitals from the list of properties.
3. Ensure that the 3D Atomistic document to be analyzed is the active document.
4. From the scrollable list, select the row(s) corresponding to the orbital(s) you wish to display.
Tip: Multiple selections are allowed. To select more than one orbital, hold down the SHIFT key
and click on the first and last orbitals in the block you want to select, or hold down the CTRL key
and click to select orbitals individually.
5. Optionally, check the View isosurface on import checkbox to automatically create an isosurface
when the potential field is imported.
6. Click the Import button.
Note: The units for the isovalues are e/Å3 and correspond to the wavefunction density, ψ2.
STM profile
CASTEP models the scanning transmission microscope (STM) profile by representing it as an isosurface
of the electron density generated only by states at a certain energy away from the Fermi level. The
distance from the Fermi level corresponds to the applied bias in actual STM experiments. Positive bias
corresponds to the imaging of empty (conduction) states, while negative bias produces images of
occupied (valence) states. This approach neglects the actual geometry of the STM tip and makes use of
the Tersoff-Hamman approximation for tunneling transport (Tersoff and Hamman, 1985).
STM profile visualization makes sense only for models that represent surfaces in the slab supercell
geometry. In addition, information about charge density at a distance from the surface is likely to be
inaccurate as a result of the DFT failing to reproduce the asymptotics of the wavefunction decay into a
vacuum.
Note: Because of the size of the .check file STM analysis is available only if orbitals are explicitly
requested.
Note: Visualization of the images created by conduction band states requires that the calculation was
performed with an increased number of empty bands. The recommended procedure is to use
electronic density of states (DOS) calculations and specify a sufficiently high number of empty states.
Tip: Field visualization in Materials Studio fully supports periodic display. You can use the Field tab on
the Display Style dialog to change the range of a field to display more or less than one unit cell of a
structure.
Note: The default settings for field visualization result in the fields being displayed over one unit cell of
a structure. Use the In-Cell display mode for the lattice (accessible from the Lattice tab of the Display
Style dialog) to make sure that the field is positioned around the displayed atoms.
energy band gap is also easily deduced from the band structure plots as it normally corresponds to the
energy difference between two states at high symmetry points.
An example of the CASTEP output for band structure is shown in Figure 1. Electronic states with different
spins are presented using different colors and the Fermi level is shown with the dashed line. All energies
are relative to the Fermi level (or to the top of the valence band in the case of insulators or
semiconductors). The labels along the X axis of the band structure graph correspond to the standard
definitions of high symmetry points for the given lattice type (Bradley and Cracknell, 1972). The Γ-point is
denoted by a G.
The density of states included in this plot was created using Interpolation as an Integration method and
Alpha and Beta as the DOS display.
Note: If you have calculated DOS for a spin polarized system with a non-zero initial spin and the total
spin is not optimized two charts will automatically be displayed, <seedname> Band
Structure Alpha.xcd and <seedname> Band Structure Beta.xcd.
Figure 1. PDOS for BN. The left window is for boron, the right for nitrogen.
To create a PDOS chart
1. Choose Modules | CASTEP | Analysis from the Materials Studio menu bar.
2. Select Density of states from the list of properties.
3. Use Results file selector to pick the right results file.
4. Select Partial to display the partial DOS.
5. Check the required angular momentum components for the PDOS.
6. Select the atom or atoms in the model, for which the PDOS is to be created.
7. For spin-polarized calculations, select the required DOS component from the DOS display
dropdown list.
8. Optionally, set additional CASTEP DOS analysis options. The Interpolation Integration method gives
more accurate results than the Smearing method, although it is slightly slower.
9. Click the View button.
10. A new chart document, seedname PDOS.xcd, is created in the results folder.
Tip: You should rename this file to indicate the atoms on which it is based. In the example above (see
Figure 1) suitable names for the two files created might be BN N PDOS.xcd and BN B
PDOS.xcd (the default names are BN PDOS.xcd and BN PDOS (2).xcd).
You can then use CASTEP to analyze the calculated stress tensors for each of these runs and generate a
file with information about elastic properties. The information in this file includes:
A summary of the input strains and calculated stresses.
Results of linear fitting for each strain pattern, including the quality of the fit.
The correspondence between calculated stresses and elastic constants for a given symmetry.
A table of elastic constants, Cij, and elastic compliances, Sij.
The derived properties such as bulk modulus and its inverse, compressibility, Young modulus and
Poisson ratios for three directions, and the Lamé constants required for modeling the material as an
isotropic medium.
If you requested the polarization property, the main output file, seedname.castep, contains the
piezoelectric tensor output at the end, for example:
Symmetrized Piezoelectric Stress Tensor (C/m^2)
-----------------------------------------------
0.000000 0.000000 0.000000 0.000000 -0.918444 0.000000
0.000000 0.000000 0.000000 -0.918444 0.000000 0.000000
-0.600778 -0.600778 1.321444 0.000000 0.000000 0.000000
Strain pattern: 1
======================
Current amplitude: 1
Transformed stress tensor (GPa) :
0.719908 -0.000000 -0.000000
-0.000000 0.331220 0.369780
-0.000000 0.369780 0.174538
Current amplitude: 2
Transformed stress tensor (GPa) :
0.115598 -0.000000 -0.000000
-0.000000 -0.019103 0.146509
-0.000000 0.146509 -0.144727
....
This provides information about the connection between components of the stress, strain, and elastic
constants tensors. At this stage, each elastic constant's representation is a single compact index rather
than by a pair of ij indices. Later, the file provides the correspondence between the compact notation
and the conventional indexing:
Stress corresponds to elastic coefficients (compact notation):
1 7 8 4 0 11
A linear fit of the stress-strain relationship for each component of the stress has the following format:
Stress Cij value of value of
index index stress strain
1 1 0.719908 -0.003000
1 1 0.115598 -0.001000
1 1 -0.509448 0.001000
1 1 -1.066455 0.003000
C (gradient) : 299.206750
Error on C : 4.882222
Correlation coeff: 0.999734
Stress intercept : -0.185099
The gradient provides the value of the elastic constant (or a linear combination of elastic constants). The
quality of the fit, indicated by the correlation coefficient, provides the statistical uncertainty of that
value. The stress intercept value is not used in further analysis. It is an indication of how far the
converged ground state was from the initial structure.
The output file then summarizes the results for all the strain patterns:
============================
Summary of elastic constants
============================
id i j Cij (GPa)
1 1 1 299.20675 +/- 4.882
3 3 3 500.96355 +/- 2.004
4 4 4 126.32975 +/- 4.509
6 6 6 229.09375 +/- 0.826
7 1 2 168.90265 +/- 4.411
8 1 3 158.33975 +/- 4.590
11 1 6 0.00000 +/- 0.000
This includes the errors only when the strain amplitude uses more than two values, since there is no
statistical uncertainty associated with fitting a straight line to only two points.
The output files present the elastic constants in a conventional 6×6 tensor form, followed by a similar
6×6 representation of the compliances:
=====================================
Elastic Stiffness Constants Cij (GPa)
=====================================
========================================
Elastic Compliance Constants Sij (1/GPa)
========================================
The elastic Debye temperature evaluation uses the recipe by Anderson (1963), where the averaged
sound velocity is evaluated numerically as an integral over all propagation directions. The Debye
temperature is proportional to the averaged sound velocity, v m:
Where:
h is the Planck constant
k is the Boltzmann constant
N is Avogadro's number
ρ is the density
M is the molecular weight of the solid
q is the number of atoms in the molecule.
Note: Elastic Debye temperature and averaged sound velocity are not reported if the calculated
elastic constants do not produce physically meaningful sound velocities in some directions (for
example, if certain Cij values or their combinations are negative when they should be positive). This
can happen for low quality calculations or for calculations that were not preceded by a geometry
optimization (including optimization of the lattice parameters).
The last section of the file contains average properties that describe the elastic response of a polycrystal,
for example:
====================================================
Elastic constants for polycrystalline material (GPa)
====================================================
Voigt Reuss Hill
Bulk modulus : 230.06015 220.35108 225.20562
Shear modulus (Lame Mu) : 137.27031 116.19120 126.73075
Lame lambda : 138.54661 142.89028 140.71845
Young modulus : 343.49337 296.46492 320.14092
Poisson ratio : 0.25116 0.27576 0.26308
This output contains the bulk modulus, shear modulus, Young modulus, and Poisson ratio averaged
according to Voigt, Reuss, and Hill schemes (Nye 1957).
In addition, a universal anisotropy index suggested by Ranganathan and Ostoja-Starzewski (2008) is
evaluated.
The Vickers hardness in GPa evaluation uses Tian's (2012) empirical formula. This expresses hardness in
terms of the polycrystalline bulk, B, and shear, G, moduli (using Pugh's modulus ratio, k=G/B):
The correlation between the results from this macroscopic model and experimentally measured
hardness values is quite good when hardness is larger than 5 GPa. However, the empirical formula
overestimates the hardness of softer materials.
Niu et al. (2019) introduced a similar empirical model to estimate fracture toughness, KIC, from the
polycrystalline moduli:
Note: The empirical expressions for Vickers hardness and for fracture toughness cannot be applied if
the calculated shear modulus is negative. This can happen when the calculation settings are not
sufficiently accurate. In such a case, the summary report does not include the hardness or the
fracture toughness.
Note: After a CASTEP job that included calculation of the electric field response (that is, the
Polarizability, IR and Raman spectra checkbox was checked on the Properties tab of the CASTEP
Calculation dialog) is complete and the results files have been downloaded successfully, Materials
Studio will automatically import the Hessian matrix, provided that the CASTEP job was not a
Properties run.
A dynamical matrix is returned if the TASK keyword in the input parameters (.param) file is set to either
Efield or Phonon+Efield and there is a Γ-point among the phonon k-vectors. You can achieve this either
by checking the Polarizability, IR and Raman spectra checkbox on the Properties tab of the CASTEP
Calculation dialog or, alternatively, if you wish to calculate the phonon dispersion or phonon density of
states, by checking the Phonons checkbox in the list on the Properties tab, checking the Dispersion or
Density of states checkboxes, as appropriate, that appear below the list, and ensuring that there is a Γ-
point in the q-vector set.
IR spectra
In order to carry out a vibrational analysis, you will need to import a Hessian matrix from the output of a
CASTEP run.
To import a Hessian and display an IR spectrum
1. Choose Modules | CASTEP | Analysis from the menu bar to display the CASTEP Analysis dialog.
2. Select IR spectrum from the list of options at the top of the dialog.
3. Ensure that the 3D Atomistic document to be updated is the active document.
4. Click the Import Hessian from CASTEP output button.
If you requested that the electric field response be calculated as part of the CASTEP job by checking the
Polarizability, IR and Raman spectra checkbox on the Properties tab of the CASTEP Calculation dialog,
you will be able to display vibrational mode frequencies and absorption intensities for the structure
being studied in the form of a list of values or graphically as a vibrational spectrum using the Vibrational
Analysis tool.
Raman spectra
To display a Raman spectrum
1. Choose Modules | CASTEP | Analysis from the menu bar to display the CASTEP Analysis dialog.
2. Select Raman spectrum from the list of options at the top of the dialog.
3. Use Results file selector to pick the right results file.
4. Ensure that the 3D Atomistic document containing the structure is the active document.
5. Select the Function of the mode to be calculated.
6. If the Function is set to Intensity set the Temperature.
7. Set the Smearing value to be used.
8. Select the Units for the X axis.
9. Specify whether to reverse the wavenumber and intensity axes.
10. Click the View button.
11. A new chart document, seedname Raman.xcd, is created in the results folder.
Note: NMR in CASTEP is part of the separately licensed module NMR CASTEP. NMR calculations can
only be performed if you have purchased this module.
Note: NMR properties are dynamic (that is, they are not present on atoms until they are assigned
during analysis).
To display the shielding tensor
1. Choose Modules | CASTEP | Analysis from the menu bar to display the CASTEP Analysis dialog.
2. Select NMR from the list of properties.
3. Make sure that the desired 3D Atomistic document is the active document.
4. Click the Assign isotropic chemical shielding to structure button to import the data into the model.
5. Right-click in the 3D Viewer and select Label from the shortcut menu. Set the Object type to Atom,
select NMRShielding from the Properties list, and click the Apply button to display the shielding
tensors.
To display the electric field gradients
1. Choose Modules | CASTEP | Analysis from the menu bar to display the CASTEP Analysis dialog.
2. Select NMR from the list of properties.
3. Make sure that the desired 3D Atomistic document is the active document.
4. Click the Assign electric field gradients to structure button to import the data into the model.
5. Right-click in the 3D Viewer and select Label from the shortcut menu. Set the Object type to Atom,
select EFGQuadrupolarCoupling or EFGAsymmetry, as appropriate, from the Properties list, and
click the Apply button to display the electric field gradients.
To display the J-coupling constants
1. Choose Modules | CASTEP | Analysis from the menu bar to display the CASTEP Analysis dialog.
2. Select NMR from the list of properties.
3. Use Results file selector to pick the right results file; in this case it should be a seedname_
NMRJC.castep file.
4. Click the Show J-coupling constants button to generate and open a study table reporting the
calculated J-coupling constants. The study table document will be named seedname_NMRJC
JCoupling.std.
The study table contains two sheets, one with the summary of the results (atom pairs,
corresponding interatomic distances, and isotropic and anisotropic components of the J-coupling
tensor) and one with the detailed breakdown of contributions to the isotropic coupling (Fermi
contact term, spin-dipole term, paramagnetic, and diamagnetic contributions).
5. Right-click in the 3D Viewer and select Label from the shortcut menu. Set the Object type to Atom,
select ServerAtomName from the Properties list, and click the Apply button to display the atom
names that are shown in the first two columns of the Summary sheet of the study table.
summary of the calculation that includes the value of the scissors operator, the geometry of the
calculation and the smearing applied. The second sheet contains a table of the real and imaginary parts
of the dielectric function versus the photon energy (eV).
It is important to set the parameters for this type of calculation correctly in order to be able to compare
the results obtained with experimental optical constants.
Scissors operator: This parameter effectively describes the difference between the theoretical and
experimental band gap values. When the experimental value is known, you can perform a band
structure calculation to find the theoretical band gap. Then simply set the scissors operator to the
difference between the two values.
Tip: When there is no experimental band gap value but there are data on the dielectric function, you
can repeat the calculations with varying scissors operator values and select the one that gives a
theoretical absorption edge which matches the experimental value.
In the absence of any experimental data, use a fixed proportion of the theoretical band gap as the
scissors operator, for example 25%. This approach allows you to study trends in optical properties
across a family of similar compounds.
Calculation: This parameter and the related Polarization or Incidence settings, specify the experimental
geometry. These are irrelevant for optically isotropic crystals but they are necessary to study anisotropy
of optical properties in crystals with lower symmetry. A good example of the theoretical study of optical
anisotropy of crystalline polymers was given by Ambros-Drachl et al. (1995). A more detailed discussion
of the meaning of these settings is given in the Optical properties theory topic.
Smearing: This parameter specifies the width of the Gaussian broadening that is used to produce the
dielectric function. It is similar to the broadening that is used in density of states calculations.
To create a dielectric function grid
1. Choose Modules | CASTEP | Analysis from the Materials Studio menu bar.
2. Select Optical properties from the list of properties.
3. Use Results file selector to pick the right results file.
4. Select the geometry of the calculation from the Calculation dropdown list.
If Polarized geometry is selected, specify the Polarization direction in the text boxes.
If Unpolarized geometry is selected, specify the Incidence direction in the text boxes.
If Polycrystalline geometry is selected, no additional parameters need to be specified.
5. Specify a Smearing value in the text box (0.5 eV is a good default setting).
6. Click the Calculate button.
7. A new grid document, seedname Optical Properties.xgd, is created in the results
folder and becomes the active document.
Display optical constants
The following optical constants can be evaluated based on the calculated dielectric function:
Reflectivity
Absorption
Refractive index, n and k
Complex optical conductivity
Energy loss function
Any of these constants and the complex dielectric function itself, can be plotted using Materials Studio.
The only additional information that might be required are the Drude correction parameters.
The empirical Drude contribution to the spectra is due to intraband excitations that are not included in
the spectra calculated by CASTEP. The Drude contribution, which may be non-zero only for metals and
semi-metals, is determined by the DC Conductivity and Drude Damping parameters. Appropriate values
for these quantities cover a wide range and are usually taken from experiment. A typical value of the
Drude Damping parameter for metals is 0.05 eV. Typical values for the DC Conductivity in metals are
around 250 eV. But it must be 0 for insulators. Another common unit for the DC conductivity is 1/
(Ohm*cm). The conversion factor between this unit and eV is 1/(Ohm*cm) = 5.924 × 10-4 eV.
To create a chart with optical properties
1. Choose Modules | CASTEP | Analysis from the Materials Studio menu bar.
2. Select Optical properties from the list of properties.
3. Ensure that the grid document, seedname Optical Properties.xgd, is active.
4. Specify the property to be plotted using the Function dropdown list.
Tip: If the system is a molecule in the box the only relevant property to display is the Absorption
calculated for Polycrystalline geometry.
5. Specify the energy units for the plot using Units dropdown list.
6. Set the Drude correction parameters for metallic systems. Click the More... button to access the
Optics Analysis Options dialog.
Tip: Ensure that the Drude term is ignored for the nonmetallic systems, that is, check that the DC
Conductivity is 0 on the CASTEP Optical Properties Options dialog.
7. Click the View button.
8. A new chart document, seedname Optical Properties.xcd, is created in the results folder. The
summary from the first sheet of the grid document is used as the graph title to assist in
identification of the graphs.
Note: The View button is enabled only if a suitable grid document is active.
Tip: If the only experimental information available is the value of the static dielectric constant, it is still
possible to assess the value of the scissors operator. Calculate the refractive index repeatedly, using
different scissors operator values and select the one that gives the best fit between the calculated n(0)
value and the square root of the experimental static dielectric constant.
Materials Studio can produce phonon dispersion and DOS charts from any .phonon CASTEP output
file. These files are hidden in the Project Explorer but a .phonon file is generated with every .castep
file that has a PhonDisp or PhonDOS suffix. All the files that can be used in the analysis of the current
calculation will be presented in the results file selector. However, it is not recommended that you create
phonon DOS plots based on the results of a phonon dispersion calculation that only uses q-vectors in
high-symmetry directions.
Tip: When evaluating phonon DOS, use only the results of the phonon calculations on the
Monkhorst-Pack grid.
To create a phonon dispersion chart
1. Choose Modules | CASTEP | Analysis from the Materials Studio menu bar.
2. Select Phonon dispersion from the list of properties.
3. Use Results file selector to pick the appropriate results file (it should have a PhonDisp suffix).
4. Select the required frequency units for the chart from the Units dropdown list.
5. Select a required display style (Points or Line) from the Graph style dropdown list.
6. Optionally, check Show DOS and set the relevant DOS options.
7. Click the View button.
8. A new chart document, seedname Phonon Dispersion.xcd, is created in the results
folder.
To create a phonon DOS chart
1. Choose Modules | CASTEP | Analysis from the Materials Studio menu bar.
2. Select Phonon density of states from the list of properties.
3. Use Results file selector to pick the right results file (it should have a PhonDOS suffix).
4. Select a required DOS type (Full or Partial) from the DOS display dropdown list.
5. Optionally, set additional CASTEP Phonon DOS analysis options. The Interpolation Integration
method gives more accurate results than the Smearing method, although it is slightly slower.
6. Click the View button.
7. A new chart document, seedname DOS.xcd, is created in the results folder when Full density of
states is selected. A document seedname PDOS.xcd, is created in the results folder when Partial
density of states is selected.
Updating structure
Updating cell parameters and atomic coordinates after Geometry Optimization or a Dynamics run
should be the first step in the Analysis process.
Note: Structure updates happen automatically after a CASTEP job is completed and the results files
have been downloaded successfully. The description below is, therefore, only relevant for jobs which
were run in standalone mode and where the results files were transferred manually, or results files in
which the final structure produced by CASTEP has been manually modified and you wish to revert to
the unmodified form.
You can update the structure if the chemical composition (the number of atomic species and the
number of atoms of each kind) has not been changed since Materials Studio saved the structure in the
results folder. After the update, the cell parameters and atom coordinates are set to the values returned
by the CASTEP run in the (hidden) binary .check file.
To update the structure
1. Choose Modules | CASTEP | Analysis from the Materials Studio menu bar.
2. Select Structure from the list of properties.
3. Ensure that the 3D model document to be updated is active.
4. Click the Update button.
Note: An error message is generated if the output .check file is incomplete, or if the chemical
composition of the model has been changed.
The Update button is enabled only if the output .check file is present and the active document is
the 3D model to be updated.
example, if a cluster node contains two 6-core chips, the value can be set to 6. This modification can
speed up parallel jobs for some architectures.
CASTEP jobs are controlled by input files that are generated by Materials Studio when you start a job.
CASTEP writes the results of the calculations to various output files, which are downloaded from the
server to the appropriate Materials Studio project when the job is completed.
CASTEP remote jobs run according to the standard sequence of processes described in A sample remote
job run with the differences explained in the Sample CASTEP run help topic.
Note: You cannot access a Remote Gateway Web Page for 3DEXPERIENCE Cloud jobs.
Below is a list of the most common reasons for CASTEP jobs to fail. It may help you to identify and fix
any problems you have with your remote CASTEP jobs. For generic reasons for remote job failures, see
If a remote job fails in the Materials Studio Online Help.
Tip: Select View | Project Log from the menu bar to see if any error or warning messages have been
reported.
Argument Description
-h Displays the help text.
-np Specifies the number of cores on which to run CASTEP. When this option is
not specified a single core is used.
number of cores The number of cores to use.
-nt Specifies desired number of threads. When this option is not specified a
single thread for each process is used.
number of The number of threads to use.
threads
-q Submits the job to the specified queue.
queue name The name of the queue on which to run the job.
seedname The seed used to identify the set of CASTEP input and output files. The input
files should be present in the directory in which the CASTEP script is started.
If you wish to calculate properties, you should execute the script with all the seednames as parameters
for the RunCastep.sh:
RunCASTEP.sh -np 4 seedname seedname_BandStr seedname_DOS seedname_Optics
Note: By default large seedname.check files are removed after the execution to save space. In
most cases the seedname.castep_bin files are enough to restart the calculation. However, if
check files are needed they can be requested by adding the following command:
-extraoptions KeepCheckFiles:Yes
For example:
RunCASTEP.sh -extraoptions KeepCheckFiles:Yes seedname
Tip: If you have transferred files into a Materials Studio project folder, but you cannot see them in
the Project Explorer, try using the Refresh button to update the Project Explorer.
* These files are hidden by default, so they will not appear in the Materials Studio Project Explorer; these
files are only visible in the Windows File Explorer if Windows on your machine is configured to show
hidden files.
associated values.
A .cell file is generated by Materials Studio when you launch a calculation or when you click the Save
Files button on the CASTEP Job Files dialog.
Under certain circumstances, you may wish to manually add keywords to a .cell file. The procedure is
described in the topic Running CASTEP in standalone mode.
The following rules govern keywords and data blocks in the .cell file:
Keywords and data blocks may appear in any order.
The keywords, data blocks, and values are case- and punctuation-insensitive.
Only one keyword and its options may appear on a line.
Data blocks are preceded by %BLOCK <keyword>, followed by a series of data lines, and
terminated by %ENDBLOCK <keyword>.
Keywords and data blocks that do not appear in the .cell file are set to their default values.
Units may be specified for physical quantities.
Note: Recognized units and their identifiers are listed in the UNIT keyword topics. If no units are
specified, the default units will be assumed.
Comment lines (those beginning with #, ;, ! or COMMENT) and blank lines are ignored.
Keywords and their values are fully described in the topic CASTEP keywords.
Values are always specified in atomic units and one of the following letter codes will be present:
E - total energy and enthalpy
T - temperature
R - atomic coordinates (Cartesian)
h - cell matrix
hv - derivatives (velocities) of the cell matrix
F - forces
V - velocities
S - stress tensor
P - pressure
Supercell approach
CASTEP is based on a supercell method, you must perform all studies on a periodic system, even when
the periodicity is superficial. For example, you must represent a crystal surface by a finite-length slab
(Figure 1). To study molecules, you must assume that they are in a box and treat them as periodic
systems. There is no limitation on the shape of the supercell.
Figure 1. Schematic representation of the supercell calculation of a small molecule adsorption on a surface
The main advantage of imposing periodic boundary conditions relates to Bloch's theorem. This states
that in a periodic system you can write each electronic wavefunction as a product of a cell-periodic part
and a wavelike part (Payne et al. 1992).
Eq. CASTEP 1
You can then expand the cell periodic part, ψ, using a basis set consisting of a discrete set of plane waves
whose wave vectors are reciprocal lattice vectors of the crystal:
Eq. CASTEP 2
Therefore, you can write each electronic function as a sum of plane waves, exp[i(k+G)•R]. For details of
the advantages and technical details of the plane wave basis set, see Plane wave basis set on page 108.
Among the major gains is the simplified form of the Kohn-Sham equations:
Eq. CASTEP 3
In this form, the kinetic energy is diagonal, and the various potentials (electron-ion, Hartree, exchange-
correlation) are described in terms of their Fourier transforms.
Another advantage of the plane wave basis set is the ease of calculating derivatives of the total energy
with respect to atomic displacements (that is, stresses and forces are computationally cheap in this
approach). This enables the implementation of efficient geometry optimization and molecular dynamics
schemes. It is also straightforward to study and improve convergence of plane-wave basis set
calculations, as there is a systematic way of adding more basis functions.
However, there is an additional cost to using the supercell approach for systems that lack periodicity in
three dimensions inherently (isolated molecules, defects in solids, surfaces, and so on). In a study of a
single defect, for example, the properties of an infinite array of defects (Figure 2) would always be
calculated. Therefore, it is essential to introduce enough separation between artificial images of such
nonperiodic objects to ensure that there is no appreciable interaction between them. The situation is
similar for surface calculations carried out in a slab geometry. Ensure that the vacuum layer is thick
enough to eliminate artificial interactions between the slabs.
The PBE functional has been designed to give essentially the same results as the PW91 functional but it
is more robust in systems with rapidly varying electron density. The RPBE functional was designed
specifically to improve the DFT description of the adsorption energies of molecules on metallic surfaces.
For a detailed comparison of these three functionals when applied to surface science problems, see
Marlo and Milman, 2000. One of the latest developments, PBEsol functional, is explicitly designed to
improve the first-principles gradient expansion for exchange over a wide range of density gradients. This
therefore improves the description of equilibrium properties of densely packed solids and their surfaces.
Implementation of all gradient-corrected functionals in CASTEP follow the method described by White
and Bird (1994). The application of gradient-corrected exchange-correlation functionals in total-energy
calculations using a plane-wave basis set is less straightforward than in localized basis set approaches
(used in other programs such as DMol3). The usual form of the exchange-correlation potential includes
gradients whose calculation requires the use of a high-quality representation of the density. This is
computationally expensive in both memory and time. These problems are overcome by defining an
exchange-correlation potential for the discrete set of grid points consistent with the discretized form of
the exchange-correlation energy used in the plane-wave total-energy calculations. This potential is
calculated exactly on the minimum fast Fourier transform grid and gives improved convergence and
stability as well as computational efficiency.
Meta-GGA functional
In addition to the generalized gradient (GGA) functionals, which depend on the local density and its
gradient, CASTEP can handle functionals that depend on the kinetic energy density. The current
supported functional is RSCAN (Bartók and Yates, 2019), which is an improved regularized version of the
SCAN ("Strongly Constrained and Appropriately Normed") functional developed by Sun et al., 2016.
Meta-GGA functionals are considered to be more accurate than pure GGA, and the cost of such
calculations is significantly lower than for nonlocal functionals.
The SCAN functional satisfies all known constraints that a semi-local functional can satisfy. The remaining
free parameters are fitted to reproduce exact or accurate reference values, or norms, of exchange and
correlation energies.
The RSCAN modification eliminates the unphysical divergence of the exchange-correlation potential,
which occurs in some free atoms. It also fixes the feature of the switching function in SCAN, which
introduces rapidly oscillating regions in the exchange-correlation potential. These improvements make
the SCAN functional usable in the context of robust pseudopotential generation. CASTEP's
implementation has shown that RSCAN is transferable and accurate for a broad range of solid state and
molecular systems.
Note: There are limitations on the use of meta-GGA in CASTEP. The meta-GGA functional is not
compatible with spin-orbit coupling, J-coupling, linear-response phonons, or polarizability
calculations. You cannot use the formalism with mixture atoms. You can only use on-the-fly
generated pseudopotentials.
Note: There is an important difference between standard DFT calculations with local exchange-
correlation potentials and the nonlocal exchange case. The potential used in the latter scenario
depends on the SCF k-points, while in the former case, the potential depends only on electron
density. This difference can make all properties calculations a lot more expensive in terms of memory
usage and CPU time. Limit the number of SCF k-points that you use in such situations. This is
particularly relevant for small unit cells, where the default settings might generate a very large k-point
set.
spin orbit interaction) is insufficient to model the electronic band structure. In magnetic systems, the
spin orbit coupling introduces a preferential direction (or set of directions) for the spin density.
The prediction of many physical properties requires the inclusion of spin orbit coupling. For example,
effective masses in semiconductors (modeling light and heavy holes), the anomalous Hall effect, Berry
curvature, topological insulators, and magnetoelectrics.
The plane waves and pseudopotential model only explicitly models the pseudized valence electrons.
These electrons have a velocity, v << c, and as such the standard (non-relativistic) Kohn-Sham (KS)
equations are appropriate. Work by Dal Corso et al. (2005) allowed the development of j-dependent
pseudopotentials so that CASTEP can include spin orbit coupling . These pseudopotentials act on KS
spin or wavefunctions and an electron density with a vector spin at each grid point.
DFT+U (LDA+U)
DFT can only give access to the ground state properties of a system, even if you know the exact
exchange-correlation potential. To model electronic excitations, CASTEP substitutes Kohn-Sham
equations with the Dyson equation, using the energy-dependent self-energy operator instead of the
unknown exchange-correlation potential. The local and energy-independent exchange-correlation
potential can represent this operator, in which case you can recover the Kohn-Sham equations for one-
electron excitations. However, this approach is not appropriate for a large class of compounds, so-called
correlated systems. These systems require a mixture of the Hubbard model theory and DFT. The most
popular implementation is the DFT+U method, as introduced by Anisimov and coworkers (1991).
The CASTEP implementation of DFT+U adopts a simplified, rotationally invariant approach (Cococcioni
and de Gironcoli, 2005, Dudarev et al., 1998). The only external parameter required for this approach is
the effective value of the on-site Coulomb parameter, U, for each affected orbital. You can calculate this
parameter theoretically (Cococcioni and de Gironcoli, 2005). However, experience shows that the best
results are achieved when the U parameter can vary depending on the property of interest. CASTEP
therefore does not calculate the value of U, but uses it as an input parameter.
Symmetry in CASTEP
You can trace back the original idea of using space group symmetry in plane-wave total energy
calculations to the k290 package (Kunc et al., 1985). The importance of this issue for achieving high
performance has long been recognized. Often a high symmetry system has a fairly small unit cell. This
means that the Brillouin zone is quite large and efficient Brillouin zone sampling requires a high number
of k-points. When you recognize a relationship between the contributions to the charge density from
wavefunctions with k-vectors related by symmetry, you can greatly reduce this number. It is then
sufficient to use only one k-point from each star of symmetry-related points. Because the cost of a
CASTEP calculation increases linearly with the number of k-points, any symmetry-related reduction in
the number of k-points gives a direct computational speedup.
CASTEP during geometry optimization also uses space group symmetry. CASTEP always symmetrizes
forces on atoms, the stress tensor, and consequently atomic displacements and changes in the cell
parameters, so that the space group symmetry can never become lower. However, CASTEP does not
perform symmetry analysis. Instead, it uses symmetry information provided by Materials Studio.
Pseudopotentials
The concept of a pseudopotential is a crucial one for plane-wave total energy methods since the
alternative full Coulomb potential of the electron-ion interaction decays too slowly to be accurately
represented by a small number of Fourier components.
Another way of looking at the problem is to analyze wavefunctions. Consider a solid as a collection of
valence electrons and ion cores. The ion cores contain nuclei and tightly bound core electrons. The
where |lm> are the spherical harmonics and Vl is the pseudopotential for angular momentum l. A
pseudopotential that uses the same potential in each angular momentum channel is called a local
pseudopotential. Local pseudopotentials are computationally much more efficient than a nonlocal ones,
however, only a few elements can be described accurately using local pseudopotentials.
An important concept in the pseudopotential applications is the degree of hardness of a
pseudopotential. A pseudopotential is considered soft when it requires a small number of Fourier
components for its accurate representation and hard otherwise. Early development of accurate norm-
conserving pseudopotentials quickly showed that the potentials for transition metals and for first row
elements (O, C, N, and so on) turn out to be extremely hard (Bachelet et al., 1982; Kerker, 1980 ). Various
schemes have been suggested to improve convergence properties of norm-conserving
pseudopotentials (Troullier and Martins, 1991). Norm-conserving potentials in CASTEP are generated
using the kinetic energy optimization scheme developed by Lin et al., (1993) and Lee (1996).
A more radical approach was suggested by Vanderbilt (1990), which involves relaxing the norm
conservation requirement in order to generate much softer pseudopotentials. In the ultrasoft
pseudopotential scheme the pseudo-wavefunctions are allowed to be as soft as possible within the core
region, so that the cutoff energy can be reduced dramatically. Ultrasoft potentials (USP) have another
advantage besides being much softer than the norm-conserving potentials. The generation algorithm
guarantees good scattering properties over a pre-specified energy range, which results in much better
transferability and accuracy of the pseudopotentials. USP usually also treats shallow core states as
valences by including multiple sets of occupied states in each angular momentum channel. This also
adds to high accuracy and transferability of the potentials, although at a price of computational
efficiency.
Transferability is the main benefit of the pseudopotential technique over all electron DFT
implementations. Pseudopotentials are constructed from a fixed electronic configuration of an isolated
atom or ion and thus they reproduce the scattering properties of a nucleus in that particular
configuration. They can also be used reliably for any other atomic configuration or in various solid-state
environments due to the way in which they are generated, which guarantees the correct scattering
properties over a wide energy range. An example of the accuracy that can be achieved across a number
of structures and chemical environments is given by Milman et al. (2000).
Nonlocal pseudopotentials, even in their most efficient separable representation (Kleinman and
Bylander, 1982), still consume a large share of the computational time in pseudopotential total-energy
calculations. In addition, the cost of applying nonlocal pseudopotentials in reciprocal space increases
with the cube of the number of atoms. Therefore, it becomes one of the dominant operations for large
systems. The nonlocality of the pseudopotential, however, extends only over the region occupied by
the core of the atom. As the core region is relatively small, especially if the system has a lot of vacuum in
it (slab calculations, open zeolite systems, and so on), it is more efficient in this case to apply the
pseudopotential in real space. The computational cost of this method increases with the square of the
number of atoms and so is preferable for big systems. CASTEP uses the scheme developed originally for
norm-conserving potentials (King-Smith et al., 1991 ) and was subsequently generalized for the case of
ultrasoft pseudopotentials.
The separation of electrons into core and valence creates a problem when treating exchange-correlation
interactions. Within the core region of the nucleus the densities of the two subsystems overlap which
makes it difficult to cleanly descreen the pseudopotential during its generation. The only term in the
potential energy operator that is nonlinear with respect to the electron density is the exchange-
correlation energy. It has been shown by Louie et al. (1982) that there is a simple explicit scheme for
treating the nonlinear exchange and correlation interaction between the core and the valence charge
densities. This approach leads to significant improvements in the transferability of the potential. In
particular, the spin-polarized calculations become noticeably more accurate. The nonlinear core
correction (NLCC) term is especially important when semicore states are not explicitly treated as
valences. On the other hand, their explicit inclusion into the valence subsystem essentially removes the
need for the NLCC treatment.
Norm-Conserving Pseudopotentials
The main requirement of the pseudopotential approach is that it reproduces the valence charge density
associated with chemical bonds. For pseudo and all-electron wavefunctions to be identical beyond the
core radius, Rc, the integrals of squared amplitudes of the two functions must be the same (Hamann et
al., 1979). This is equivalent to requiring norm-conservation from pseudo wavefunctions, that is, each
should carry exactly one electron. This condition ensures correct reproduction of the scattering
properties of the pseudopotential.
Figure 1. Schematic representation of the all-electron and pseudized wavefunctions and potentials
The typical method for generating pseudopotentials is as follows:
All-electron calculations are carried out for an isolated atom in a chosen electronic configuration (not
necessarily in the ground state). This provides valence electron eigenvalues and valence electron
wavefunctions for the atom (shown as ψ in Figure 1).
Selection of a parameterized form for the ionic pseudopotential (or the pseudo wavefunction).
The parameters are then adjusted, so that a pseudoatom calculation with the same exchange-
correlation potential as in the all-electron atom gives:
Pseudowavefunctions, ψps (Figure 1), that match the valence wavefunctions outside some cutoff
radius, Rc.
Pseudoeigenvalues that are equal to the valence eigenvalues.
This procedure involves direct inversion of the radial Kohn-Sham equation for parameterization of the
pseudowavefunctions rather than the pseudopotential. Normalizing each wavefunction, pseudo and all-
electron, to one, results in automatically satisfaction of the norm-conservation constraint because of
matching the wavefunctions outside Rc.
The ionic pseudopotentials construction uses Rc ranging from one to two times the value of the physical
core radius. The smaller the value of Rc, the harder and more transferable the potential. The conflicting
effect of Rc creates an obvious trade-off between accuracy and efficiency.
Optimized Pseudopotentials
A number of recipes exist for producing pseudopotentials that are optimized with respect to the energy
cutoff required in solid-state calculations.
Lin et al. (1993) suggested the following generation scheme based on the earlier work of Rappé et al.
(1990).
Where:
jl(qir) are spherical Bessel functions with (i-1) zeros between r=0 and r=Rc. The cutoff radius value
is as large as possible, consistent with satisfactory transferability of the potential.
2. Minimization of the kinetic energy beyond the cutoff vector qc determines the coefficients αi:
Eq. CASTEP 6
where qc is chosen to be equal to q4 from Eq. CASTEP 5. Lagrange multipliers satisfy some additional
constraints:
The normalization of the pseudowavefunction.
The continuity of the first two derivatives of the pseudowavefunction at Rc.
3. The standard step of inverting the radial Kohn-Sham equation produces a smooth pseudopotential
with optimal convergence properties.
Lee (1996) suggested further enhancements and it is essentially this scheme that generated the bulk of
norm-conserving pseudopotentials in the CASTEP database. This generation method eliminates the
condition that the second derivative of the pseudo wavefunction must be continuous at a given cutoff
radius because it automatically satisfies the second derivative constraint. The scheme allows you to tune
qc for a given Rc to optimize the accuracy and efficiency of a pseudopotential.
The USP implementation allows you to run CASTEP calculations with a lower energy cutoff than their
NCP counterparts; producing a clear advantage in terms of the calculation time. However, USP
formalism is more complex and becomes close to intractable for such complex concepts as linear
response implementation for phonons or NMR properties, or for nonlocal exchange-correlation
functionals. There is a number of tasks and properties that CASTEP can address only with norm-
conserving potentials. The database of tabulated NCPs available in Materials Studio is rather old and in
some cases might contain potentials that have not been tested sufficiently. A recommended alternative
is to use on-the-fly generated norm-conserving pseudopotentials that are more accurate and much
better tested.
Ultrasoft pseudopotentials
Ultrasoft pseudopotentials (USP) were introduced by Vanderbilt (1990) in order to allow calculations to
be performed with the lowest possible cutoff energy for the plane-wave basis set. Since it is understood
that there are inherent limits on optimizing the convergence of norm-conserving pseudopotentials, a
completely different approach has been designed.
The rationale behind USP is that, in most cases, a high cutoff energy is only required for the plane-wave
basis set when there are tightly bound orbitals that have a substantial fraction of their weight inside the
core region of the atom. In these situations, the only way to reduce the basis set is to violate the norm-
conservation condition by removing the charge associated with these orbitals from the core region. The
pseudo wavefunctions are thus allowed to be as soft as possible within the core, yielding a dramatic
reduction in the cutoff energy.
Technically, this is accomplished by introducing a generalized orthonormality condition. In order to
recover the full electronic charge, the electron density given by the square moduli of the wavefunctions
is augmented in the core regions. The electron density can thus be subdivided into a soft part extending
through the unit cell and a hard part localized in the core regions.
USP formalism in solid state
The total energy in the USP scheme has the same form as other pseudopotential plane-wave methods,
with the nonlocal potential VNL given as:
Eq. CASTEP 7
where the projectors β and coefficients D(0) characterize the pseudopotential and differ for different
atomic species. The index I refers to an atomic site. The electron density in the expression for the total
energy is given by:
Eq. CASTEP 8
where ϕ are the wavefunctions and Q(r) are the augmentation functions that are strictly localized in the
core regions. The ultrasoft potential is fully determined by the local part, Vlocion(r) and by the
coefficients D(0), Q, and β. The algorithm that generates these quantities is described below.
The relaxation of the norm-conserving constraint is achieved by introducing a generalized
orthonormality condition:
Eq. CASTEP 9
where coefficients q are obtained by integrating Q(r). Kohn-Sham equations can be rewritten for the
USP case as:
Eq. CASTEP 11
where Veff contains Vlocion(r), Hartree and exchange-correlation terms. All the terms arising from the
augmented part of the electron density are grouped with the nonlocal part of the pseudopotential (Eq.
CASTEP 7) by defining new coefficients:
Eq. CASTEP 13
The difference compared with the norm-conserving case is the presence of the overlap operator S, the
wave-function dependence of the coefficients D and the fact that the number of projector function β is
typically twice as large. A number of calculations related to augmentation charges can be carried out in
real space because of the localized nature of the function involved, so that the overall impact of these
extra steps on efficiency is not large.
Further technical details of the USP implementation, including expressions for the derivatives of the
total energy, can be found in Laasonen et al. (1993).
Pseudopotential generation
As in the norm-conserving case, an all-electron calculation is carried out on a free atom, leading to a
screened atomic potential, VAE(r). For each angular momentum a set of reference energies, εl, is
chosen. Typically two reference energies are sufficient. These energies should cover the range over
which good scattering properties are required. At each reference energy the solution of the radial Kohn-
Sham equation which is regular at the origin is obtained. Next cutoff radii Rcl are chosen and for each all-
electron wavefunction ψ generated above, a pseudo wavefunction ϕ is constructed. The only constraint
is that it should join smoothly to ψ at Rcl. A smooth local potential, Vloc(r), is generated to match VAE(r)
smoothly at a cutoff radius, Rloc. An auxiliary radius, R, is defined to be slightly higher than all of the core
radii. Finally, the local orbitals are formed (they vanish beyond R):
Eq. CASTEP 14
It is now possible to define the quantities needed for the solid state calculation (Vlocion(r), D(0), Q, and
β):
Eq. CASTEP 15
Eq. CASTEP 16
Eq. CASTEP 18
Note: The non-linear core correction term can be introduced in the descreening procedure, Eq.
CASTEP 17, in exactly the same way as in the norm-conserving case.
Note: The formalism described above is based on the non-relativistic Schroedinger approach. With
CASTEP you can also use a scalar-relativistic approach either as a Koelling-Harmon (1977) scheme or as
its approximation commonly referred to as ZORA, zeroth order regular approximation (Lenthe et al.
1994).
The following conditions are satisfied in this scheme:
The pseudo eigenvalues are equal to the all-electron ones, and the corresponding orbitals match
exactly outside the core radii.
The scattering properties are correct at each reference energy. Thus, the transferability can be
systematically improved by increasing the number of such energies.
The valence charge density is precisely equal to the all-electron valence density in the reference
configuration.
Nonlinear core correction
The nonlinear core correction (NLCC) was first proposed by Louie et al. (1982) as a way of obtaining
accurate pseudopotential descriptions of magnetic systems. It has been shown, however, that NLCC is
equally important for spin non-polarized systems with semicore electrons.
The reason that the NLCC is required is found in the total energy expression in DFT:
Eq. CASTEP 19
In the pseudopotential formalism, the charge density is divided into core and valence contributions. The
energy of the core is assumed to be constant and is subtracted out. Thus, in Eq. CASTEP 19 above the
total charge density is replaced by a (pseudo) valence charge density and Eion is evaluated using
pseudopotentials.
As a result all interaction between core and valence electrons is transferred to the pseudopotential. This
implies linearization with respect to charge density which can only be an approximation to the kinetic
energy and the explicitly nonlinear exchange-correlation energy.
This linearization is obviously a good approximation when core and valence densities are well separated
in space. However, if there is significant overlap between the two densities, it will be the source of
systematic errors and will lead to reduced transferability of the potential.
The solution to the NLCC problem lies in a modifying the pseudopotential generation procedure and the
solid-state algorithm. In a typical generation scheme a screened atomic potential Vl is produced for each
angular momentum channel subject to certain constraints, for example norm-conservation, matching
eigenvalues to all-electron values, and so on.
These screened potentials give rise to atomic pseudowavefunctions that represent valence states only.
The valence pseudo charge density is constructed from these wavefunctions and is used to descreen the
potential to produce the bare ionic potential:
Eq. CASTEP 20
Since the exchange-correlation potential is a nonlinear function of the charge density, it can be shown
that this procedure, especially for spin-polarized systems, creates ionic pseudopotentials that are
dependent on the valence configuration. Louie et al. (1982) suggested substituting Eq. CASTEP 20 with
the following expression:
Eq. CASTEP 21
Here the total exchange-correlation potential is subtracted from the screened atomic potential. In
addition, the core charge must now be added to the valence charge whenever the exchange-correlation
potential is calculated. This additional atomic information has to be passed on to CASTEP and the core
charge is deemed to be the same in all applications. A drawback of this is that the core charge is difficult
to represent accurately on the Fourier grid used in pseudopotential calculations. Furthermore, the core
density is usually bigger than the valence density, so its presence might mask effects related to valence
charge density.
The next step in developing the formalism was to introduce the partial core correction. This approach
recognizes the fact that only the area of overlap between the core and valence densities is of interest.
Core density close to the nucleus is of no consequence although this is precisely the part of the core
density that causes the problems described above. Partial NLCC replaces the full core density with a
function that coincides with ρc outside a certain radius, and falls off smoothly toward the nucleus.
CASTEP uses a partial core correction with numerical core density in the case of ultrasoft
pseudopotentials for certain elements. The database of norm-conserving potentials does not, however,
include NLCC information.
Real-space representations of pseudopotentials
The nonlocality of the pseudopotential only extends over the region occupied by the core of the atom.
As the core region is relatively small, it should be possible to deal efficiently with the nonlocality of the
pseudopotential by working in real space. Only a small number of operations should be required to
project the angular momentum components of each wavefunction in the core of each atom.
Furthermore, the number of operations needed to project the angular momentum components of a
single wavefunction around a single atom in real space is independent of the size of the system, leading
to better scaling with system size than in the reciprocal space projection scheme. The computational
cost of the real-space implementation increases with the square of the number of atoms, compared
with the cubic relationship that exists for reciprocal space projection methods (Payne et al., 1992).
CASTEP uses the scheme developed originally for norm-conserving potentials (King-Smith et al., 1991)
and subsequently generalized for the case of ultrasoft pseudopotentials.
where Vl(r) is the original nonlocal projector and jl(qr) is the spherical Bessel function.
Select a real-space cutoff radius, R0. Generally this is about 1.5-2.0 × Rc, where Rc is the core radius
used in pseudopotential generation.
Variationally select χl(q) in the intermediate region, Gmax < q < γ by minimizing:
Eq. CASTEP 23
The object of this step is to achieve the best possible convergence of the modified potential in real
space.
Transform χl(q) back to real space to get χl(r ) which is now used in the real-space evaluation of
nonlocal potential matrix elements in solid-state calculations.
The modified real-space projectors are no longer identically zero outside Rc, although the minimization
in Eq. CASTEP 23 ensures the best possible convergence compatible with the removal of the wrap
around error. A quantitative measure of the error introduced by the transformation is given by:
Eq. CASTEP 24
This error function should be small (less than 0.001) for all wavevectors, q, and can be inspected in the
CASTEP output file.
One should bear in mind that the use of real-space scheme represents an approximation and that more
accurate results are obtained with the reciprocal-space version. The accuracy of the real-space
transformation is controlled by the value of the cutoff radius, R0. The default set of these values in
CASTEP is designed to be for accuracy rather than performance. It is advisable to carry out independent
tests on small systems before embarking on large-scale calculations with drastically reduced R0 values.
The real space transformation of the pseudopotential depends on the cutoff energy for plane wave basis
set, and thus it is not possible to precalculate real-space projectors in the same way as the reciprocal
space potentials (which are provided in CASTEP database).
cutoff energy as illustrated in Figure 1 (the radius of the sphere is proportional to the square root of the
cutoff energy).
The truncation of the basis set at a finite cutoff energy will lead to an error in the computed total energy
and its derivatives. It is possible to reduce the magnitude of the error in a systematic way by increasing
the value of the cutoff energy. In principle, the cutoff energy should be increased until the calculated
total energy converges within the required tolerance. This would be the recommended procedure if, for
example, you were conducting a phase stability study where the absolute values of total energies of
different structures are compared. However, the energy differences for the same structure converge at
much lower cutoff energies than the total energy itself. This is due to the cancellation of errors and
allows one to use moderate basis sets for reliable geometry optimization or molecular dynamics studies.
These arguments are also valid for the convergence testing of Brillouin zone sampling.
Finite basis set correction
One of the difficulties associated with the use of plane-wave basis sets is that the number of basis states
changes discontinuously with cutoff energy. In general these discontinuities will occur at different
cutoffs for different k-points in the k-point set. In addition, at a fixed cutoff energy, a change in the size
or shape of the unit cell will cause discontinuities in the plane wave basis set. This problem can be
reduced by using a denser k-point set, so that the weight attached to any particular plane wave basis
state is reduced. However, the problem is still present even with quite dense k-point samplings. It can be
handled approximately by applying a correction factor that accounts for the difference between the
number of a states in a basis set with an infinitely large number of k-points and the number of basis
states actually used in the calculation (Francis and Payne, 1990).
The finite basis set correction is of the utmost importance when cell optimization is performed with a
basis set that is not absolutely converged. For example, the norm-conserving silicon pseudopotential is
sufficiently soft and provides accurate results with a cutoff energy for plane waves of about 200 eV.
However, if one calculates the equation of state using this cutoff (that is, the volume dependence of the
total energy and pressure), the volume that corresponds to the minimum in total energy does not
coincide with the volume that gives zero pressure. As one repeats such EOS calculations at progressively
higher cutoff energies and with better k-point sampling, the difference between the two volume values
becomes smaller and smaller. In addition, the E-V curve calculated at low cutoff energy is jagged but
becomes more and more smooth as the cutoff energy is increased. The reason for the jagged
appearance of the E-V curve is the discontinuous change in the number of plane waves used at the same
cutoff energy but for different lattice constants. The finite basis set correction, when added to the total
energy, allows one to perform calculations with a fixed number of basis states and to interpolate the
results as if the more physical condition of a fixed energy cutoff were used. A detailed description of the
recommended use of this technique can be found in Milman et al. (1994).
The only parameter that must be known in order for this correction term to be evaluated is dEtot/d
lnEcut, where Etot is the total energy of the system and Ecut is the cutoff energy. CASTEP can calculate
this term automatically or you can input this parameter manually.
The value of the dEtot/d lnEcut gives a good indication of the convergence of the calculation with respect
to the energy cutoff and k-point sampling. When its value per atom (that is, the value reported in the
CASTEP output file divided by the number of atoms) is smaller than 0.01 eV/atom, the calculation can be
considered very well converged. A value of 0.1 eV/atom is sufficient for most calculations.
Preconditioning
All electronic relaxation methods in CASTEP use preconditioning to speed up the rate of convergence
(Payne et al., 1992). The idea is to find a matrix which, when multiplied with the residual vector, gives the
exact error in the wavefunction. Formally, this matrix can be written down as:
Eq. CASTEP 28
where εn is the exact eigenvalue for the band being optimized. It in not possible to evaluate this matrix,
but by recognizing that the kinetic energy dominates the Hamiltonian for large G-vectors, it is possible to
approximate the matrix by a diagonal function which converges to G-2 for large G-vectors and is
constant for small G-vectors. The actual form of the matrix is (Payne et al., 1992):
Eq. CASTEP 29
where x is the ratio of |k + G|2 to the kinetic energy of the given state.
Application of the preconditioning scheme causes all of the large wave-vector components of a
wavefunction to converge at nearly the same rate.
Density mixing schemes
CASTEP supports four schemes for generating new starting charge densities from the results of the
previous SCF step (Kresse and Furthmuller, 1996). The simplest is the linear mixing scheme. This method
requires one parameter and the output density is produced as a linear combination of the input and
output densities on the last iteration.
Kerker mixing recognizes the fact that small-G components of the charge density should be mixed with
smaller weights in order to prevent charge sloshing during SCF optimization. The new density is given
by:
Eq. CASTEP 30
The Kerker scheme is, consequently, defined by two parameters, the mixing amplitude, A, and the
cutoff wavevector, Gmax.
The most efficient mixing scheme is Pulay's scheme. In this approach, the input charge density and the
residual vectors are stored over a number of SCF optimization steps. A new input density is obtained in
each step as a linear combination of charge densities of all previous steps. The new density is
determined such that it minimizes the normal of the residual vector subject to the constraint of
conserving the number of electrons. Pulay's scheme starts with a Kerker-type step, so it is effectively
controlled by three parameters: the mixing amplitude, A, the cutoff wavevector, Gmax, and the depth of
the optimization history.
Broyden mixing is very similar to the Pulay scheme, except that it performs a quasi-Newton relaxation
instead of DIIS minimization of the residuals.
Treatment of metallic systems
The main difference between metals and insulators, from the technical point of view, is that the number
of occupied bands is not the same at different k-points in the Brillouin zone. The number of occupied
bands for insulators is calculated as one half the total number of valence electrons, but this approach is
not suitable for metals. Partial occupancies are introduced to eliminate discontinuous changes in total
energy that are created when an energy band crosses a Fermi level during SCF minimization.
The overall strategy for treating metallic systems in CASTEP is as follows. The number of bands has to be
slightly higher than would be required for an insulator. It may be necessary to add more bands if
convergence is slow, especially for spin-polarized systems. An artificial electronic temperature is then
introduced by assuming Gaussian-like smearing of each energy level (de Vita, 1992). Occupation
numbers are deduced from the ratio of the area of the Gaussian that falls below the Fermi level to its
total area. Thus, a level deep in the valence band has an occupation number of 1, a level directly on the
Fermi surface has an occupation number of 0.5, and a level high above the Fermi energy is empty. In
some sense, this is the same as giving the Kohn-Sham particles a finite temperature. Therefore, CASTEP
uses the Mermin finite temperature expression for the total energy (Marzari, 1997):
E = EKS - TS
where:
T is the temperature (that is, smearing width)
S is the Mermin entropy due to the different ways in which the partially filled bands could have been
filled
The CASTEP output is as follows. The line
Final energy, E = -861.8315373731 eV
gives the Kohn-Sham energy, EKS, which ignores the entropy. However, it is the total Mermin free
energy that is minimized by CASTEP and reported in the SCF cycles in the line
Final free energy (E-TS) = -861.8315374521 eV
If you were actually interested in the electronic states at some reasonable nonzero temperature, you
would smear the Fermi level with a thermal (that is, Fermi-Dirac) smearing method and the free energy
reported by CASTEP would be the value you require.
Note: Temperature equivalent of energy is given by the conversion factor of 11604.5 K/eV; so the
default smearing width of 0.1 eV corresponds to a very high temperature of 1160 K.
However, it is often the T = 0 K value that is required and the smearing is necessary to improve the
convergence of the numerical algorithm. As the smearing ('temperature') is increased, the TS term
increases and the free energy, EKS - TS, goes down. At the same time, as the entropy contributes more
and more to the free energy, the electronic minimization pays less attention to theEKS term, so the final
EKS increases. Because the dependence of the free energy and the Kohn-Sham energy on the smearing
width is the same (to first order), averaging the two values gives a much better approximation to the T =
0 K result:
NB est. 0K energy (E-0.5TS) = -861.8315374126 eV
The total energy calculated by CASTEP for metals is thereby corrected for the fact that it now includes
the artificial electronic entropy contribution. This correction is possible because a closed analytical form
for the dependence of the total energy on the smearing width (or electronic temperature), s, exists. In
principle, evaluation of Etot (s -> 0) is necessary in order to obtain a physically meaningful energy. In the
past, values for s as low as 0.01 eV have been used to calculate the converged energy. The total energy
correction described above allows the use of a much higher smearing width (up to 1 eV), while still giving
results converged with respect to s. The advantage is the added stability of the calculation, since
occasional Fermi-level crossings do not create instabilities. However, there is no simple expression for a
similar correction for either atomic forces or stress on the unit cell.
The error in the forces introduced by this limitation is generally small and acceptable. Nonetheless,
geometry optimizations for metals should be performed with some caution. You should use smaller
smearing values than for single-point energy runs. The calculated forces will be more accurate, but there
is a risk of instability, since changes in atomic geometry are likely to cause reordering of bands and some
Fermi-level crossings. Cell optimization is also more difficult for metallic systems, since the stress tensor
could be affected by nonphysical contributions from nearly empty conduction bands.
Disorder in solids
Many crystal structures possess static positional disorder. The crystal may either contain fewer atomic
species than there are crystallographically equivalent sites to populate, or there might, for example, be
statistical occupancy of a given crystallographic site by more than one type of atom in different unit
cells. This positional disorder is typically manifested in site occupancy factors of less than unity for the
average unit cell.
Materials Studio allows you to model positional disorder by specifying atom occupancy and by defining
mixture atoms.
Atom occupancy
The vast majority of atoms will have an occupancy of unity. However, in the case of an atom on a
partially occupied interstitial position, you can specify an occupancy ranging from 0.0 to 1.0.
Atom occupancies can be modified in the Properties Explorer.
Mixture atoms
Atomic sites in a crystal can also be described in terms of a hybrid atom that consists of two or more
element types. The relative concentrations can be set for any number of atoms, but the total
concentration must not exceed 100%. The mixture atoms description is the most often used
representation of solid solutions, metallic alloys, disordered minerals, and so on.
The basic ideas of the workable VCA implementation for DFT methods, in particular for the
pseudopotential-based techniques, were set out by Bellaiche and Vanderbilt (2000). The main statement
for the implementation of VCA with ultrasoft potentials can be expressed as:
Eq. CASTEP 86
where the total external potential is generated as the sum of the nonlocal potentials of each atomic
species, α, taken with the weights, w, of the component atoms in the mixture atom. This means that all
the key components of the ultrasoft potentials, that is, the local part, Vloc, and the D and Q matrices,
are all weighted according to the site occupancies.
The VCA scheme described above has been applied successfully to study disorder in the perovskites Pb
(ZrTi)O3 (Bellaiche and Vanderbilt, 2000) and BiScO3-PbTiO3 (Iniguez et al., 2000), and Al-Si disorder in
silicates such as hollandite and gehlenite (Winkler et al., 2002).
Note: Ramer and Rappé (2000) suggested an alternative scheme where the pseudopotentials are
generated on the fly for the mixture atoms. Their scheme appears to be less readily generalized for
the case of more than two disordered species or where the mixture atom's components have
different chemical valences.
where one part of the mixture is an atom of "nothing". This case is not supported for a number of
technical and fundamental reasons.
Current implementation of the molecular dynamics algorithms in CASTEP is not yet compatible with the
VCA formalism. None of the MD schemes (NVE, NPT, and so on) works for disordered crystals in the
current version.
Linear response calculations (phonons, polarizability) are not supported for disordered crystals.
Test results
The VCA implementation in CASTEP is capable of correctly describing a number of mixture atoms, as
shown by the results below.
Bulk modulus of NaCl/KCl solid solution
The bulk modulus of the halite-sylvite mixture (NaCl/KCl) has been calculated as a function of
concentration by calculating the P-V equation of state for different Na concentration and fitting them
with the third-order Birch-Murnaghan analytical expression. The results agree with the experimental
data (Figure 1) to better accuracy than is typically achieved for the ordered crystals. The norm-
conserving pseudopotentials were used in this study (all other tests described below were carried out
with ultrasoft pseudopotentials).
Figure 1. Dependence of the bulk modulus of the NaCl/KCl solid solution on concentration.
Experimental data are from Walker et al. (2004). The solid line shows a polynomial fit to the calculated results.
Table 2. Structure of the R3m rhombohedral phase of PZTO. Experimental data are from Amin et al.
(1981), earlier VCA results are from Ramer and Rappé (2000).
a (Å) α (°) Ox Oy Pbx Tix Pb-O (Å) Ti-O (Å)
CASTEP 4.0749 89.535 -0.0040 0.5080 0.5586 0.0250 2.516 1.973, 2.115
exp. 4.0809 89.676 -0.0046 0.5092 0.5620 0.0216 2.537 1.994, 2.098
VCA 4.0215 90 (fixed) - - - - - -
Table 3. Structure of the R3c rhombohedral phase of PZTO. Experimental data are from Jirak and
Kala (1988), earlier VCA results are from Ramer and Rappé (2000).
a (Å) α (°) Ox Oy Oz Pbx Tix Pb-O Ti-O (Å)
(Å)
CASTEP 5.7726 59.589 0.4942 0.0284 -0.0180 0.0283 0.2632 2.565 1.979,
2.111
exp. 5.8240 59.600 0.4860 0.0330 -0.0190 0.0341 0.2620 2.487 2.024,
2.109
VCA 5.6790 60 - - - - - - -
(fixed)
Table 4. Energy differences between the three phases of PZTO, in meV per 5 atom cell. Earlier VCA
results are from Ramer and Rappé (2000).
E(R3m) - E(P4mm) E(R3c) - E(P4mm)
CASTEP -4.7 -6.3
VCA 9 6
These results indicate that the structural features are reproduced quite accurately within VCA. Note that
the experimental data for rhombohedral phases are for Zr concentrations slightly different from 0.5,
since these phases are not stable for the 1:1 composition. The energy differences between the phases
are extremely small in recent studies and earlier work by Ramer and Rappé (2000). DFT energies cannot
be trusted at this level of accuracy to judge the relative stability of the polymorphs. For example, the
fact that the angles were fixed to 90 and 60° in the earlier VCA work suggests that after the full
optimization, the rhombohedral phases are likely to become more stable than the tetrahedral phase, in
agreement with the CASTEP results (Table 4).
The structure of leucite
Tetragonal leucite, KAlSi2O6, contains three different sites with a fully disordered distribution of Al and Si
cations. The structure is only slightly distorted with respect to the cubic cell and the accuracy of
reproduction of this distortion is a measure of the quality of the VCA description. The results presented
below (Table 5) show exceptionally good agreement with the experimental data. Furthermore, atomic
positions are also predicted very accurately. The highest deviation from the experimental data for any
individual fractional coordinate is 0.001 in the CASTEP calculations, compared with maximum deviations
of up to 0.007 in the forcefield modeling.
Table 5. Lattice parameters of leucite compared to experimental data (Mazzi et al., 1976 ) and the
results of forcefield calculations (Winkler et al., 1991).
a (Å) c (Å) c/a
CASTEP 13.1497 13.8318 1.052
exp. 13.09 13.75 1.050
empirical 12.9884 13.8000 1.062
Geometry optimization
CASTEP supports a number of schemes for geometry optimization: BFGS, TPSD, and damped molecular
dynamics.
shape, although the minimizer disregards this dependence. When using Fixed Basis Size modifies the cell
geometry, the effective cutoff energy changes but the number of plane waves remains fixed. If this
change takes the dEtot/d(ln Ecut) function far away from the point used to evaluate the finite basis set
correction, the results are not accurate. Therefore, compare the starting and final geometries and
perform a completely new run starting from the final configuration if the difference between the two
structures is large.
Constraints
CASTEP can perform geometry optimization with constraints applied. The simplest type of constraint is
to fix the atom positions. In CASTEP, this means fixing the fractional coordinates of the atoms.
Note: You do not need to manually fix atoms on special positions as the symmetry treatment in
CASTEP ensures the correct behavior of such atoms.
It is also possible to fix lattice parameters. You do not need to fix lattice parameters, except for
parameters not fixed by symmetry. This type of constraint is often useful in phase stability studies
when, for example, you want to fix the angles to 90° (Winkler and Milman, 1997).
It is also possible to impose more general linear constraints on atomic coordinates. A matrix that
transforms the Cartesian coordinates of all atoms to the subspace of unconstrained coordinates
specifies these constraints. You can use this facility to fix, for example, the z-coordinates of atoms in slab
calculations of surface processes.
Non-linear constraints
Non-linear constraints refer to constraints on interatomic distances (bonds), angles, and torsions. You
can impose such constraints using the delocalized internals optimizer.
Estimated compressibility
You can facilitate cell optimization convergence by making a reasonable initial estimate of the bulk
modulus for the material. The initial changes to lattice vectors are calculated on the basis of this value,
based on the calculated stress tensor. When the estimated bulk modulus is too high, the steps become
smaller than is ideal for the minimizer. Therefore, it is conceivably possible that you could reduce the
number of geometry optimization steps by reducing this value. Conversely, too small a bulk modulus
for a hard material could result in an oscillating behavior, where the first step is too big. Thus, prior
knowledge of the hardness of the material could help to reduce the number of steps required for
calculations. The table below illustrates the effect of a judicious choice of the initial bulk modulus value.
The times are relative to the same calculation using an initial value of 500 GPa.
molecular dynamics run. A full description of the method implemented in CASTEP can be found
elsewhere (Probert, 2003).
Dynamics
Molecular dynamics (MD) involves the stepwise integration of Newton's equations from a given starting
point. It is the most natural method of performing equilibrium statistical-mechanical calculations using a
simulation. Additionally, MD may be used to explore conformations, such as molecular adsorption sites
on a surface. If frictional forces are added to the equations of motion, Langevin dynamics becomes
possible, as does geometry optimization using damped MD. Finally, extended Lagrangian methods
allow thermostatic control of the temperature.
Molecular dynamics in total-energy DFT schemes is implemented in essentially the same way as in
conventional forcefield based methods. The main difference is that the energy and atomic forces are
derived by solving DFT equations rather than from empirical potentials of interatomic interactions.
Electrons are kept on the Born-Oppenheimer surface by means of explicit electronic structure
optimization after each MD step. A side effect of this is that evaluation of force and energy from first
principles is always the most computationally expensive part of ab initio MD. As a result, the efficiency of
the MD step itself has no impact on the speed of the calculation.
MD in CASTEP is based on the velocity Verlet algorithm for integration of the equation of motion. The
default time step can be as high as 1 fs with this algorithm because of its numerical stability and the
conservation of the constant of motion.
The efficiency of the method is due in large part to the introduction of wavefunction extrapolation
between MD steps (Arias et al., 1992). In addition, charge density extrapolation (Alfe, 1999) has been
introduced.
The coefficients α and β are calculated by determining the best quadratic extrapolation of the atomic
positions. This extrapolation scheme not only speeds up the calculation but also improves conservation
of the constant of motion and dramatically reduces the "electronic drag" effect (Payne et al., 1992).
An improvement suggested by Alfe (1999) recognizes the fact that the charge density is mostly
determined by the superposition of the atomic charge densities. Therefore one can write:
Eq. CASTEP 33
where ρ(t) is the self-consistent charge density at time t, and ρat(t) is the sum of the atomic charge
densities. The charge at time t+dt is simply extrapolated as:
Eq. CASTEP 34
The atomic charge density can be calculated exactly and cheaply using the same methods as for
Mulliken population analysis. The resultant scheme includes extrapolation of wavefunctions using Eq.
CASTEP 31, extrapolation of charge density difference using Eq. CASTEP 35 and recalculation of the
atomic charge densities.
Experience shows that the second-order extrapolation scheme described above is not necessarily the
best or the fastest algorithm. CASTEP therefore offers a selection of extrapolation schemes that differ in
the way the coefficients are selected: 1st order extrapolation with variable coefficients; 2nd order
extrapolation with variable coefficients; alternating first and second order extrapolation steps with
variable coefficients; and similar combinations of 1st and 2nd order schemes with fixed coefficients. The
actual setting is controlled by the values of the and MD_EXTRAP_FIT keywords in the .param file. The
fixed coefficient scheme is based essentially on a Taylor expansion, while the variable coefficients
schemes are based on the best fit between extrapolated and MD-predicted atomic coordinates. It has
been shown that the accuracy of wavefunction extrapolation improves greatly when the coefficients are
chosen based on the extrapolation of atomic coordinates.
An alternative to using wavefunction extrapolation is to use extended Lagrangian formulation, XL-BOMD
(Steneteg 2010, Niklasson 2011) . The XL-BOMD scheme allows simultaneous propagation of nuclear and
electronic degrees of freedom and shows good energy conservation at time steps that are unusually
high for first principles calculations.
Ensembles
Integrating Newton's equations of motion allows you to explore the constant-energy surface of a
system. However, most natural phenomena occur under conditions where the system is exposed to
external pressure and/or exchanges heat with the environment. Under these conditions, the total
energy of the system is no longer conserved and extended forms of molecular dynamics are required.
Several methods are available for controlling temperature and pressure. Depending on which state
variables (the energy E, enthalpy H (that is, E +PV), number of particles N, pressure P, stress S,
temperature T, and volume V) are kept fixed, different statistical ensembles can be generated. A variety
of structural, energetic, and dynamic properties can then be calculated from the averages or the
fluctuations of these quantities over the ensemble generated.
Both isothermal (where heat is exchanged with a temperature bath to maintain a constant
thermodynamic temperature) and adiabatic (where no heat exchange occurs) ensembles are available:
Table 1. Thermodynamic ensembles handled by CASTEP
Ensemble
Constant temperature, constant pressure (NPT)
Constant energy, constant volume (NVE)
Constant pressure, constant enthalpy (NPH)
Constant temperature, constant volume (NVT)
NVE ensemble
The constant-energy, constant-volume ensemble (NVE), also known as the microcanonical ensemble, is
obtained by solving the standard Newton equation without any temperature and pressure control.
Energy is conserved when this (adiabatic) ensemble is generated. However, because of rounding and
truncation errors during the integration process, there is always a slight fluctuation, or drift in energy.
True constant-energy conditions (that is, without temperature control) are not recommended for the
equilibration phase of the simulation because, without the energy flow facilitated by temperature
control, the desired temperature cannot be achieved.
However, during the data collection phase, if you are interested in exploring the constant-energy surface
of the conformational space, or if for some other reason you do not want the perturbation introduced
by temperature- and pressure-bath coupling, this is a useful ensemble.
Nosé NVT dynamics
The Nosé thermostat (Nosé, 1984) generates deterministic dynamics, with the temperature controlled
by a fictitious additional coordinate, s, added to the Lagrangian of the system. The thermostat employs
a feedback loop between the instantaneous kinetic energy and the required temperature. The rate of
feedback is determined by the mass parameter, Q. This parameter should be chosen so that the natural
oscillation frequency of the Nosé coordinate is close to the characteristic frequency of the actual system.
The mass parameter is related to the thermostat relaxation time by:
Eq. CASTEP 36
Where:
g is the number of degrees of freedom (usually 3N - 3, where N is the number of atoms)
k B is the Boltzmann constant
T is the thermostat temperature
τ is the relaxation time.
An improvement on the standard Nosé thermostat (or Nosé-Hoover thermostat) is the Nosé-Hoover
chain method, described in detail by Tuckerman et al. (2001). In this method the kinetic energy
fluctuations of the thermostat variable are controlled by coupling it to another thermostat variable. The
kinetic energy fluctuations of the second thermostat are, in turn, controlled by coupling to a third
thermostat, and so on, to form a chain of M thermostats. This new coupling (the Nosé-Hoover "chain"
thermostatting mechanism) leads to a more general canonical dynamics method.
The Lagrangians of the system are modified such that a term representing the kinetic energy of the cell
depends on a user-defined mass-like parameter W. An elastic energy term pΩ is related to the pressure
and volume or the stress and strain of the system. The equations of motion for the atoms and cell
vectors can be derived from this Lagrangian. The motion of the cell vectors, which determines the cell
shape and size, is driven by the difference between the target and internal stress.
With hydrostatic pressure only:
Eq. CASTEP 37
Where:
h={a•b•c} is the cell vector matrix
G=h'h
r i=hsi
ϕ is the interaction potential
Ω is the volume of the cell
Tr is the trace of a matrix
The dots above some symbols indicate time derivatives and the primes indicate matrix transposition.
With stress, the elastic term pΩ is replaced by:
Eq. CASTEP 38
Where:
S is the applied stress
Ω0 is the initial volume
ε is the strain
ε is also a tensor, defined as:
Note: It is the user-defined variable W, which determines the rate of change of the volume/shape
matrix.
A large W means a heavy, slow cell. In the limiting case, infinite W reverts to constant-volume dynamics.
A small W means fast motion of the cell vectors. Although that may mean that the target stress can be
reached faster, there may not be enough time for equilibration.
NPH ensemble
The constant-pressure, constant-enthalpy ensemble (Andersen, 1980) is the analogue of constant-
volume, constant-energy ensemble, where the size of the unit cell is allowed to vary.
Constraints
CASTEP supports linear constraints defined using the matrix Aij:
Eq. CASTEP 39
The summations are over Ni ions and Nc constraints. Vector Ri is the Cartesian position of the ith atom
and C is a constant. These constraints can be used to, for example:
Fix some of the Cartesian coordinates of the atom
Fix the x-separation between a pair of atoms
Constrain an atom to lie in a plane
Constrain the center of mass to be fixed
Example
The example below (Figure 1) shows the results of an NVE calculation for a nitrogen diatomic molecule
with using a time step of 1 fs. The energy conservation is excellent, the energy drift over 500 steps
(0.5 ps) is less than 0.1 kcal/mole. The power spectrum of the trajectory gives the vibrational frequency
as 2340 cm-1, which is within 1% of the experimental value of 2360 cm-1.
Figure 1. Bond length of the N 2 molecule and power spectrum of the trajectory
In the case of anisotropic systems, angular-dependent experiments enable the separation of states with
different symmetries for the involved orbitals. An important consequence is that symmetry states,
which result solely from chemical bonding, can be studied.
Emission spectroscopy
X-ray emission spectroscopy (XES) is a classical technique used to study the electronic structure of bulk
samples. It is ideally suited to these systems due to the large information depth (around 0.1 m for soft X-
rays). In XES the core hole created by an X-ray absorption process is filled by the decay of a valence
electron. For this process energy is conserved by the emission of an x-ray photon of matching energy. A
selective, resonant excitation of a core hole state at the atomic center of an adsorbate can be used to
probe surfaces.
Absorption spectroscopy
Electron energy-loss near-edge structure (ELNES or EELS) and x-ray absorption near-edge structure
(XANES) spectroscopies provide information on the unoccupied electronic states by exciting one
electron from a core level into an unoccupied state. Scanning transmission electron microscopes with an
electron energy-loss spectrometer attached have been used extensively in the study of nanoscale
materials, providing information on local composition, structure, and electronic structure. However, the
probing electron or x-ray interacts with a system that has been perturbed by the probe itself. The
system being studied has had an inner-shell electron removed, leaving a core hole.
Energy broadening
The ELNES calculation can be performed with an energy resolution far better than that used to obtain
experimental results. Therefore, the calculated ELNES must be broadened before comparing it with the
experiment. The broadening observed in the experimental spectra has three origins: the lifetime of the
core hole, the lifetime of the excited state, and instrumental broadening. A good description of this
subject was given by Hebert (2007).
Instrumental broadening is easiest to simulate: it can be modeled by a Gaussian with a FWHM given by
the instrument. For a TEM equipped with a field emission gun, this is usually about 0.6-0.7 eV. Some
instruments are equipped with a monochromator to improve energy resolution to about 0.1 eV. In
nearly all cases the energy distribution can be modeled by a Gaussian curve.
The broadening by the core hole lifetime depends on the atom and the edge. Values for lifetime
broadening are tabulated and can be found in the literature, see Fuggle and Inglesfield (1992).
Broadening due to the lifetime of the core hole is, however, relatively low for edges accessible in EELS
(about 0.2 eV for the Ni or Cu L3 edge). The broadening of core hole lifetimes can be modeled by a
Lorentzian (Fuggle and Inglesfield, 1992).
The broadening by the excited state lifetimes is more complicated. At the edge onset (where the lifetime
of the excited state approaches infinity) the broadening is zero and increases with increasing energy.
Different methods have been employed to obtain the broadening factor, see Hebert (2007) for a
comparative review. An empirical linear function Γ(E) = 0.1 E (where E is the energy above the threshold)
gives good results when compared to the experiment - this is the option employed in CASTEP.
Spin-orbit splitting
Experimental spectra record transitions involving core states that are split as a result of the spin-orbit
interactions; so one should distinguish between states such as p 1/2 and p 3/2 and so on. The CASTEP
implementation lacks spin-orbit terms, so the calculations can only produce "generic" spectra for p, d,
or f states. In spectroscopic terms this means producing (for example) L2,3 spectra rather than just L2.
The simplest, albeit slightly approximate, way to model experimental spectra is to modify the calculated
spectrum as follows:
Find the experimental value of the spin-orbit splitting of the core states.
Generate an approximate L3 spectrum by shifting the L2 spectrum.
Apply intensity scaling based on the occupancy of the respective electronic shell. The intensity ratios
for L spectra are 2:4 and for M spectra 4:6, and so on.
Add the two shifted and scaled spectra together.
For a more general discussion of the core level spectroscopy and details of its implementation see S.-P.
Gao et al. (2008) and references.
branches in the phonon spectrum. The contribution to the partial density of states on atom i, from each
phonon band, is evaluated using:
Eq. CASTEP 41
where e j is the eigenvector (normalized to unit length) associated with the mode of energy Ej. The
projected density of states is then obtained by summation of these contributions over all phonon
bands. By construction all the projected phonon DOS sum up to the true phonon DOS.
The Pauli repulsion between two like-spin electrons, described by the size of D(r), is taken as a measure
of the electron localization. Using the corresponding factor found for a uniform electron gas, Dh(r),
Becke and Edgecombe, defined ELF as:
Eq. CASTEP 88
where
Eq. CASTEP 89
Becke and Edgecombe attributed the ratio χBE(r) to a dimensionless localization index, calibrated with
respect to the uniform electron gas. This reference was chosen arbitrarily, so ELF is only a relative
measure of the electron localization. High ELF values show that at the examined position the electrons
are more localized than in a uniform electron gas of the same density. For example, when η(r) = 1/2 the
effect of the Pauli repulsion is the same as in a uniform electron gas of the same density.
The ELF formulation originates from the Hartree-Fock description of electron gas. A formulation for ELF
in DFT has been developed by Savin et al. (1992). This DFT-ELF approach evaluates the Pauli kinetic
energy density based on the diagonal elements of the one-particle density matrix (the electron density).
The Pauli kinetic energy is the energy due to the redistribution of the electrons in accordance with the
Pauli principle and is the integral of the Pauli kinetic energy density:
Eq. CASTEP 90
The Pauli kinetic energy density itself does not resolve the bonding situation. It is an arbitrary division of
tP(r) by the kinetic energy density of a uniform electron gas of the same electron density, this yields all
the information:
is formally identical to the ratio χBE(r) in the HF approximation. The same applies for an open shell
system, where the kinetic energy densities are computed for the corresponding spin part only. So the
ELF formulas based on χBE(r)) and χS(r) are identical. In the interpretation of Savin and coworkers (1992)
ELF is a measure of the influence of Pauli principle as given by the Pauli kinetic energy density, relative to
a uniform electron gas of the same density. Similarly, for the original definition ELF does not mirror tP
(r).
Introduction
A disadvantage of the use of a plane wave (PW) basis set is that, because of the delocalized nature of the
basis states, it provides no information regarding the localization of the electrons in the system. In
contrast, a Linear Combination of Atomic Orbitals (LCAO) basis set provides a natural way of specifying
quantities such as atomic charge, bond population, charge transfer, and so on.
CASTEP performs population analysis using a projection of the PW states onto a localized basis using a
technique described by Sanchez-Portal et al. (1995). Population analysis of the resulting projected states
uses the Mulliken formalism (Mulliken, 1955). This technique is widely used in the analysis of electronic
structure calculations performed with LCAO basis sets.
Formalism
The eigenstates |ψα(k)>, obtained from the PW calculation when sampling at a given wavevector k, are
projected onto Bloch functions formed from an LCAO basis set |ϕμ(k)>. In general, such a localized basis
set is neither orthonormal nor complete. Therefore, care must be taken when performing the
projection.
The overlap matrix of the localized basis set, S(k), is defined as:
Eq. CASTEP 63
Where:
Nα is the number of PW states.
wk are the weights associated with the calculated k-points in the Brillouin zone.
p(k) is the projection operator of Bloch functions with wavevector k generated by the atomic basis set:
Eq. CASTEP 65
Where |ϕμ(k)> are the duals of the LCAO basis, such that:
Eq. CASTEP 66
The spilling parameter varies between one, when the LCAO basis is orthogonal to the PW states, and
zero, when the projected wavefunctions perfectly represent the PW states.
The density operator may be defined:
Eq. CASTEP 67
Where:
nα are the occupancies of the PW states.
|χα(k)> are the projected PW states p(k)|ψα(k)>.
|χα(k)> are the duals of these states.
From this operator, you can calculated the density matrix for the atomic states as follows:
Eq. CASTEP 68
The density matrix P(k) and the overlap matrix S(k) are sufficient to perform population analysis of the
electronic distribution. In Mulliken analysis (Mulliken, 1955) the charge associated with a given atom, A,
is determined by:
Eq. CASTEP 69
The weight of a band on a given orbital can be calculated by projecting the band onto the selected
orbital. So, the weight of band α on orbital μ is given by:
Eq. CASTEP 71
The contribution of each band α to the density of states is multiplied by the weight Wαμ(k) to get the
projected density of states for orbital μ.
Implementation details
The elements of the overlap matrix S(k) are computed in reciprocal space, as the imposition of periodic
boundary conditions implies that the orbitals need only be evaluated on a discrete reciprocal space grid.
The overlaps between orbitals on different atomic sites are calculated on the same grid with the
application of a phase factor.
The overlaps between the plane wave states and the basis functions are also calculated in reciprocal
space as this is the natural representation of the plane wave states:
Eq. CASTEP 72
Eq. CASTEP 73
Eq. CASTEP 77
For spin dependent systems, the populations of the two spin components are calculated separately and
the sum and difference of the atomic and overlap populations are calculated to find the net charge and
spin, respectively.
Note: Normally CASTEP reports a "charge spilling parameter", obtained from summation of σ (see Eq
CASTEP 78) over all occupied bands. Additionally, CASTEP can calculate an "all bands spilling
parameter", obtained from the sum over all bands, including unoccupied states. Its value is typically
higher and it describes the accuracy of the representation of conduction states through LCAO-type
functions.
In addition to providing an objective criterion for bonding between atoms, the overlap population can
be used to assess the covalent or ionic nature of a bond. A high value of the bond population indicates a
covalent bond, while a low value indicates an ionic interaction. You can obtain a further measure of ionic
character from the effective ionic valence, defined as the difference between the formal ionic and
Mulliken charges on the anion species. A value of zero indicates a perfectly ionic bond, while values
greater than zero indicate increasing levels of covalency. See Segall et al. (1996b) for more details.
Take care when interpreting the bond order values for small unit cells. The results are incorrect
whenever an atom is bonded to a number of periodic images of itself. For example, bond orders for the
primitive cell of silicon are reported as 3.0, while the same calculation for the conventional cell gives 0.75.
The factor of four appears because CASTEP sums the bond orders for the four bonds that link the Si
atom at the origin its periodic images, in the primitive cell. A manual analysis of the multiplicities is
required in such cases, specifically: determine the number of bonds to periodic images of the same atom
and divide the bond order by that number.
Note: The physical meaning of Mulliken charges or bond populations for metallic systems is unclear.
This calculation is allowed, but the scientific interpretation of the results is left to the user.
The shielding tensor, σ(r), connects the induced magnetic field to the applied magnetic field:
Eq. CASTEP 43
where Q is the projection operator onto the unoccupied Hilbert subspace. The system is solved by a
conjugate gradient minimization scheme, whereby the tedious summation over unoccupied states is
avoided. The method is, thus, comparable in computational cost to the calculation of the ground-state
electronic structure.
It follows from Table 1 that a grid of at least 12 × 12 × 12 is required to achieve convergence of the
chemical shift to within 1-2 ppm, and finer grids might be needed to achieve higher accuracy.
NMR properties are very sensitive to atomic positions, which makes NMR such a useful experimental
tool for structure analysis. This implies that it is highly recommended to perform a geometry
optimization run prior to the NMR calculation. If the structure of a crystal or molecule under
investigation is obtained from an experimental study, it is advisable to optimize positions of hydrogen
atoms first of all. If the forces on heavy atoms as reported by CASTEP are high (in excess of 1 eV/Å) then
complete structure optimization is recommended.
NMR J-couplings
NMR J-coupling or nuclear spin-spin coupling is an indirect interaction of the nuclear magnetic moments
mediated by the bonding electrons. It is manifested as the fine structure in NMR spectra, providing a
direct measure of bond strength and a map of the connectivities of the system.
The J-coupling mechanism is an essential component of many NMR experiments. Joyce et al. (2007)
developed a method to calculate J-coupling constants from first-principles in extended systems in a
planewave-pseudopotential density functional theory framework, using the projected augmented wave
method (PAW) to reconstruct the all-electron properties of the system. This method has been validated
for a small number of systems containing light atoms against quantum chemical calculations and against
experimental data (see for example, Joyce et al., 2008). This technique has been further developed by
Green and Yates (2014) to incorporate scalar relativistic effects in the ZORA (zeroth-order regular
approximation) approach and hence to provide a highly efficient method for predicting J-coupling in
extended systems containing heavy ions at negligible extra computational cost compared to the non-
relativistic method.
A general review of the formalism is given by Yates (2010), this discusses various aspects of the
calculations which should be taken into account when comparing results with solid-state NMR
experiments including anisotropy and orientation of the J tensors, the reduced coupling constant, and
the relationship between the J-coupling and crystal structure.
EPR G-tensor
Electron Spin Resonance (EPR - also known as Electron Paramagnetic Resonance) is a powerful
experimental probe of the structure of defects in solids. EPR spectra of spin ½ centers are made up of
two contributions:
the hyperfine parameters
the g-tensor
The g-tensor arises from the interaction of the electronic spin with external magnetic field. This term
plays a somewhat similar role to the shielding in NMR; induced electronic currents in the sample modify
the g tensor from its vacuum value.
EPR spectrum can be modeled using the following effective Hamiltonian, bilinear in the total electron
spin S, and the applied uniform magnetic field or nuclear spins, B and II, respectively:
where:
atomic units are used
α is the fine structure constant
the summation I runs over the nuclei
The tensors AI are the hyperfine parameters, and the tensor g is the EPR g-tensor.
The Gauge Including Projector Augmented Waves (GIPAW) approach has been used to compute g
tensors in several crystalline materials including defects in α-quartz and zirconia.
The hyperfine tensor can be obtained from the ground state charge density. In CASTEP it is computed
using the PAW approach of Van de Walle and Blöchl (1993). The calculation of the g-tensor is more
involved, requiring the linear response currents. These are computed using the GIPAW approach of
Pickard and Mauri (2001). The total g-tensor is obtained as:
Where the contributions to the total change in the g-tensor are the electron Zeeman kinetic energy,
spin-orbit, and the spin-other-orbit correction. The calculation of these terms follows the approach of
Pickard and Mauri (2002) extended for the use of ultrasoft potentials by Yates et al. (2007). The GIPAW
approach has been used to compute g-tensors in several crystalline materials including defects in quartz
and zirconia by Pietrucci et al. (2006).
Optical properties
Definition of optical constants
CASTEP can calculate the optical properties of solids that are due to electronic transitions.
In general, the difference in the propagation of an electromagnetic wave through vacuum and some
other material can be described by a complex refractive index, N:
Eq. CASTEP 48
In vacuum N is real, and equal to unity. For transparent materials it is purely real, the imaginary part
being related to the absorption coefficient by:
Eq. CASTEP 49
The absorption coefficient indicates the fraction of energy lost by the wave when it passes through the
material, so that the intensity at the distance x from the surface is:
However, when performing calculations of optical properties it is common to evaluate the complex
dielectric constant and then express other properties in terms of it. The complex dielectric constant, ε
(ω), is given by:
Eq. CASTEP 51
and hence the relation between the real and imaginary parts of the refractive index and dielectric
constant is:
Eq. CASTEP 52
Another frequently used quantity for expressing optical properties is the optical conductivity, σ(ω):
Eq. CASTEP 53
Optical conductivity is usually used to characterize metals, however CASTEP is aimed more toward the
optical properties of insulators and semiconductors. The main difference between the two is that
intraband transitions play important role in the IR part of the optical spectra of metals and these
transitions are not considered at all in CASTEP.
A further property that can be calculated from the complex dielectric constant is the energy loss
function. It describes the energy lost by an electron passing through a homogeneous dielectric material
and is given by:
Eq. CASTEP 54
DFT optics
In this approach excited states are represented as unoccupied Kohn-Sham states. Transitions between
occupied and unoccupied states are caused by the electric field of the photon (the magnetic field effect is
weaker by a factor of v/c). When these excitations are collective they are known as plasmons (which are
most easily observed by the passing of a fast electron through the system rather than a photon, in a
technique known as Electron Energy Loss Spectroscopy, described by Eq. CASTEP 54 , since transverse
photons do not excite longitudinal plasmons). When the transitions are independent they are known as
single particle excitations. The spectra resulting from these excitations can be thought of as a joint
density of states between the valence and conduction bands, weighted by appropriate matrix elements
(introducing selection rules).
where u is the vector defining the polarization of the incident electric field.
This expression is similar to Fermi's Golden rule for time dependent perturbations, and ε2(ω) can be
thought of as detailing the real transitions between occupied and unoccupied electronic states. Since
the dielectric constant describes a causal response, the real and imaginary parts are linked by a Kramers-
Kronig transform. This transform is used to obtain the real part of the dielectric function, ε1(ω).
Ultrasoft pseudopotentials produce an additional contribution to optical matrix elements that is also
included in CASTEP results.
Drude correction
The intraband contribution to the optical properties affects mainly the low energy infrared part of the
spectra. It can be described sufficiently accurately using a semiempirical Drude term in the optical
conductivity:
Eq. CASTEP 57
in terms of the plasma frequency ωP and damping parameter γD which depends on many details of the
material and is usually obtained from experiment. The Drude damping parameter describes the
broadening of the spectra due to effects not included in the calculation. Examples of processes that
contribute to this broadening are electron-electron scattering (including Auger processes), electron-
phonon scattering, and electron-defect scattering. This last contribution is usually the most important.
As a result an a priori determination of the broadening would require knowledge as to the
concentrations and kinds of defects present in the sample under study.
Polarization
For materials that do not display full cubic symmetry, the optical properties will display some
anisotropy. This can be included in the calculations by taking the polarization of the electromagnetic
field into account. As mentioned above, the unit vector u defines the polarization direction of the
electric field. When evaluating the dielectric constant there are three options:
polarized: requires a vector to define the direction of the electric field vector for the light at normal
incidence to the crystal.
unpolarized: requires a vector to define the direction of propagation of incident light at normal
incidence to the crystal. The electric field vector is taken as an average over the plane perpendicular
to this direction.
polycrystal: no direction need be specified, the electric field vectors are taken as a fully isotropic
average.
Scissors operator
As discussed below, the relative position of the conduction to valence bands is erroneous when the
Kohn-Sham eigenvalues are used. In an attempt to fix this problem, inherent in DFT, a rigid shift of the
conduction band with respect to the valence band is allowed. This procedure of artificially increasing the
band gap is know as a scissor operator.
Other limitations
The nonlocal nature of the GGA exchange-correlation functionals is not taken into account when
evaluating the matrix elements but it is expected that this will have a small effect on the calculated
spectra.
Phonons and their optical effects have been neglected.
Finally, there is an intrinsic error in the matrix elements for optical transition due to the fact that
pseudowavefunctions have been used (that is they deviate from the true wavefunctions in the core
region). However, the selection rules will not be changed when going from pseudo- to all-electron
wavefunctions.
Figure 1. Convergence of the dielectric function with the number of conduction bands
Energy cutoff
The value of the energy cutoff that was used in the original SCF calculation of the ground state electron
density is an important factor that determines the accuracy of the calculated optical properties. A bigger
basis set provides more accurate self-consistent charge density and more variational freedom when
searching for wavefunctions of unoccupied states. Figure 2 shows that while it is important to converge
the calculation with respect to the planewave cutoff to obtain the correct energies for the spectral
features, the shape of those features is reached rapidly.
Figure 3. Convergence of the dielectric function with the number of k-points in the SCF calculation
Figure 4. Convergence of the dielectric function with the number of k-points in the optical matrix elements
calculation
TD-DFT optics
TD-DFT is a formally exact treatment of electronic excited states. The implementation in CASTEP follows
the work of Hutter (2003), which takes a linear response approach to computing excitation energies
directly. A benefit of this method is that the response wavefunctions for excited states are calculated
and can then be used to derive further properties such as optical matrix elements and atomic forces. In
the case of atomic forces, a geometry optimization or molecular dynamics run can be performed for a
chosen excitation.
Hutter reformulated the equations of time-dependent Hartree-Fock in the Tamm-Dancoff
approximation applied to TD-DFT (Hirata and Head-Gordon, 1999) so that they could be efficiently
implemented in a plane-wave basis set. This self-consistent eigenvalue problem is solved in CASTEP by
the user’s choice of conjugate gradient or block-Davidson minimizers. The result is a set of electronic
excitation energies and linear response wavefunctions.
A technical report on the implementation details of the TD-DFT minimizers can be found at
http://www.hector.ac.uk/cse/distributedcse/reports/castep02/.
Here, we assume the polarization is instantaneous and lossless, satisfying the Kramers-Kronig
constraints. The linear polarizability χ(1) is related to the dielectric permittivity, discussed above.
In principle, the second order polarizability χ(2) is non-zero for any material without inversion symmetry
and gives rise to 3-wave-mixing phenomena. It is a rank-3 tensor quantity related to the third derivative
of the total energy with respect to electric field. These third order derivative quantities are computed in
CASTEP using the 2n+1 theorem of the Density Functional Perturbation Theory following the method of
Miwa (2011), also used for DFPT Raman calculations. The CASTEP implementation includes neither the
phonon contribution, nor the frequency-dependence for energies above the bandgap, which modify the
response at infrared and ultraviolet regions of the spectrum respectively, and therefore is appropriate
for the visible optical range.
For a material where the susceptibility is dispersionless, the Kleinman symmetry condition is satisfied.
This allows a contraction of the 27-element χ(2) tensor into an 18-element tensor, d, that is symmetric in
its last two indices:
Depending on the space group of the crystal, there are at most 10 symmetry-distinct elements of the d
matrix.
Measured quantities are most often reported in picometers per Volt, (pm/V). The book by Boyd (2003)
contains more detailed background and discussion.
Phonons
The basic theory of phonons, or lattice vibrations, in crystals is well understood and has been described
in detail in text books such as those by Born and Huang (1954) and Ashcroft and Mermin (1976). The
importance of the phonon interpretation of lattice dynamics is illustrated by the large number of
physical properties that can be understood in terms of phonons: infrared, Raman, and neutron
scattering spectra; specific heat, thermal expansion, and heat conduction; electron-phonon interactions
and, thus, resistivity and superconductivity, and so on.
A number of approximations are made in order to arrive at the phonon description of lattice vibrations.
First of all, it is assumed that the mean equilibrium position of each ion, i, is a Bravais lattice site, Ri.
Secondly, it is assumed that the amplitude of atomic displacements is small compared to interatomic
distances. This leads to a harmonic approximation which is sufficiently accurate to describe most of the
lattice dynamical effect of interest. However, further refinement, in the form of anharmonic theory, is
required to explain physical properties at elevated temperatures when the harmonic approximation
breaks down.
The N-ion harmonic crystal can be considered as 3N independent harmonic oscillators, whose
frequencies are those of 3N classical normal modes. Thus, at each wave vector, q, in the Brillouin zone
there are 3N vibrational modes. There are two equivalent descriptions of this system, either in terms of
normal modes or phonons. Phonons are the quanta of the ionic displacement field that describe
classical sound; they are similar to photons, the quanta of the radiation field that describe classical light.
The properties of phonons can be described using a harmonic approximation based on the knowledge
of just one fundamental quantity, the force constants matrix:
Eq. CASTEP 58
Here u refers to the displacement of a given atom and E is the total energy in the harmonic
approximation. This force constants matrix (or Hessian matrix) can also be represented in reciprocal
space and the result is commonly referred to as the dynamical matrix:
Eq. CASTEP 59
Classical equations of motion can be written in the language of dynamical matrices, as an eigenvalue
problem. Each atomic displacement is described in the form of plane waves:
Eq. CASTEP 60
where the polarization vector of each mode, ε, is an eigenvector with the dimension of 3N of the
eigenvalue problem:
Eq. CASTEP 61
The dependence of the frequency, ω, on the wave vector is known as the phonon dispersion.
Each lattice always has three acoustic modes. Their dispersion relation at small values (near the Γ point)
is linear, which is characteristic of sound waves. The rest of the vibrational modes, 3N - 3, in a Bravais
lattice with basis are known as optical modes because the long wavelength optical modes in ionic
crystals can interact with electromagnetic radiation and are therefore responsible for much of the
characteristic optical behavior of such crystals. The frequencies of optical modes are finite at q = 0. A
simplistic characterization of acoustic and optical branches (Ashcroft and Mermin, 1976) would be: All
ions in a primitive cell move essentially in phase in an acoustic mode and the dynamics is determined by
the interaction between cells; an optical mode is one in which the ions within each cell exhibit a
molecular vibration mode which is broadened into a band of frequencies due to intercellular
interactions.
A popular way of classifying lattice vibrations is based on the relationship between the orientation of the
polarization vector, ε, and the propagation direction, q. In an isotropic medium for the case of a
monatomic primitive cell it is always possible to choose three solutions for a given q so that one branch
(longitudinal) is polarized along the propagation direction and the other two branches (transverse) are
polarized perpendicular to this direction. This simple relationship does not hold for an arbitrary q-vector
in an anisotropic crystal, nor for a Bravais lattice with basis. It is still common to classify vibrational
modes along high symmetry directions in simple lattices based on the transverse and longitudinal
description.
Long wavelength vibrations in polar materials deserve a special mention. The long range character of the
Coulomb forces gives rise to macroscopic electric fields for longitudinal optic (LO) phonons at the Γ-point
in polar semiconductors and insulators. The coupling between longitudinal phonons and the
nonperiodic electric field gives rise to LO-TO splitting at the Γ-point. The value of the splitting is
determined by the static dielectric constant of the crystal and by the Born effective charge of the ions.
The Born effective charge tensor of the sth ion is the partial derivative of the macroscopic polarization
with respect to a periodic displacement of all the periodic images of that ion at zero macroscopic electric
field. A good review of the practical methods of dealing with this effect is given by Baroni et al., (2001).
Thermodynamic properties of crystals can be evaluated in a fairly straightforward way based on a
knowledge of the phonon frequencies across the Brillouin zone (see Ashcroft and Mermin, 1976; Baroni
et al., 2001). These quasiharmonic results are quite accurate at moderate temperatures and only small
corrections to calculated phase transition properties can be found at temperatures as high as 1000-
2000 K (Kern et al., 1999).
The supercell method involves perturbing the positions of the atoms slightly and calculating the reaction
forces (Ackland et al., 1997). It is necessary to use supercells of the original cell when interatomic
interaction in the system is long ranged. The main advantage of this method (and of the closely related
frozen phonon technique) is that there is no need for a new formalism; any total energy scheme like
CASTEP can be used to evaluate the forces at a number of carefully selected distorted configurations.
The original frozen phonon scheme requires a displacement with the given wave vector and has been
successfully used since the early 1980s (Yin and Cohen, 1982; Ho et al., 1984). The force constants matrix
evaluation in this formalism has been used to calculate interplanar force constants (Wei and Chou, 1994)
and, thus, phonon dispersion along high symmetry directions. More recent applications are based on
the full reconstruction of the force constants matrix (Ackland et al., 1997; Parlinski et al., 1997; and
references in Baroni et al., 2001).
Linear response calculations seek to evaluate the dynamical matrix directly for a set of q-vectors. The
starting point of the linear response approach is evaluation of the second-order change in the total
energy induced by atomic displacements. The main advantage of the scheme is that there is no need to
artificially increase the cell size in order to accommodate small values of the q-vectors, as in the frozen
phonon method, or to overcome the long range interaction problem (force constants matrix from
supercell calculations). A more detailed description of the linear response method can be found in
Baroni et al., (2001). The CASTEP implementation is described in the Linear response topic.
Interpolation schemes
The force constant matrix (Eq. CASTEP 58) for non-polar insulators and most metals decays
proportionally to r-5 or even faster. This fact allows the introduction of an effective interpolation scheme
to accelerate phonon calculations within both the finite displacement and linear response formalisms
through the introduction of a cutoff radius, Rc, beyond which all the elements of the force constant
matrix can be treated as zero. If an approximation for the force constant is available, the dynamical
matrix for any given reciprocal vector can be calculated using Eq. CASTEP 59.
Finite displacement: supercell method
In the supercell method, finite displacement calculations are carried out in the supercell, circumventing
the sphere described by the cutoff radius, which directly give all the non-zero elements of the force
constant matrix.
Linear response: Fourier interpolation
In the Fourier interpolation method, calculations are performed on the regular reciprocal mesh (with a
spacing corresponding to the real space cutoff radius). The set of dynamical matrices obtained in these
calculations can be used to construct the force constant matrix using Fourier transformation.
Infrared spectra
In the infrared energy range (typically below 1 eV), vibrations are an important component of optical
spectra. In the case of metals, this process is screened considerably by the plasmon (intraband)
electronic contribution, while for semiconductors and insulators, it is vibrations that are mainly
responsible for absorbing light energy.
Theoretically, infrared absorption intensities are described in terms of a dynamical matrix (also known as
a Hessian) and Born effective charges (also known as atomic polarizability tensors, ATP). The Born
effective charge tensor of an ion is the partial derivative of the macroscopic polarization with respect to a
periodic displacement of all the periodic images of that ion at zero macroscopic electric field. A good
review of the practical methods of dealing with this effect is given by Baroni et al. (2001). The Born
effective charge tensor is calculated within the linear response formalism by applying a Gonze
approximation. Once these data are known, it is straightforward to calculate IR intensities (see
Calculating vibrational frequencies). IR absorption intensities are printed in the CASTEP output file for
any run where phonons at the Γ-point and LO-TO splitting are calculated.
The concept of IR absorption intensities is mostly used in the analysis of molecules. CASTEP can be used
to model isolated molecules if the molecule in question is assumed to be contained within a box (see
Setting up a calculation on an isolated molecule. This approach allows accurate prediction of IR
absorption intensities. Figure 1 compares the results of such calculations obtained using DMol3 and
CASTEP for a variety of molecules.
Figure 1. Comparison of IR spectra calculated using DMol3 and CASTEP for a range of molecules
Linear response
Linear response, or density functional perturbation theory (DFPT), is one of the most popular methods
of ab initio calculation of lattice dynamics (Baroni et al. 2001). However, the applicability of the method
extends beyond the study of vibrational properties. Linear response provides an analytical way of
computing the second derivative of the total energy with respect to a given perturbation. Depending on
the nature of this perturbation, a number of properties can be calculated. A perturbation in ionic
positions gives the dynamical matrix and phonons; in magnetic field - NMR response; in unit cell vectors
- elastic constants; in an electric field - dielectric response, and so on.
The force constants matrix can be obtained by differentiating the Hellmann-Feynman forces on atoms,
with respect to ionic coordinates. This procedure reveals that the force constants matrix depends on the
ground state electron charge density and on its linear response to a distortion of atomic positions. Due
to the variational principle of the density functional formalism, the second order change in energy
depends on the first order change in the electron density.
The DFPT formalism is, in many ways, very similar to the density functional theory (DFT) itself. DFT states
that the total energy is the functional of the electron density; thus one can solve the DFT equations by
minimizing the total energy. Similarly, the DFPT problem can be solved by minimizing the second order
perturbation in the total energy, which gives the first order changes in density, wavefunctions and
potential (Gonze et al. 1994, Gonze 1997).
This variational technique for solving DFPT problems is a more robust and accurate method than the
original Green's function, iterative scheme suggested by Baroni et al. 1987. Both methods should
converge to the same solution but the Green's function method does not rely on the variational
principle and its convergence properties are thus inferior. Moreover, once the first order change in
wavefunctions, ψ(1), is evaluated, then the Green's function scheme evaluates the second order energy,
which is not stationary with respect to first-order errors in ψ(1). Thus, the variational scheme of Gonze
1997 is expected to be more accurate.
The electronic second order energy, which is minimized in this approach, as implemented in CASTEP, is:
Eq. CASTEP 62
where the superscripts refer to the ground state (0) and first (1) and second (2) order changes,
respectively. Similar terms have to be evaluated for the ionic terms in the total energy. The
preconditioned conjugate gradients minimization scheme can be used to find the minimum of this
functional with respect to the first order wavefunctions. The dynamical matrix for a given q is then
evaluated from the converged first order wavefunctions and densities (Gonze 1997).
Raman spectra
Raman spectroscopy is used to study the vibrational, rotational, and other low-frequency modes in a
system. It is based on the Raman effect of inelastic scattering of monochromatic light. This interaction
with vibrations results in the energy of incident photons being shifted up or down. The energy shift is
defined by the vibrational frequency and the proportion of the inelastically scattered light is defined by
the spatial derivatives of the macroscopic polarization, technical details are described by Porezag and
Pederson (1996).
Spatial derivatives of the macroscopic polarization are calculated numerically along eigenvectors of each
Raman active phonon mode by calculating the polarization for each displacement using a linear
response formalism. Once these derivatives are known, it is straightforward to calculate the Raman
cross-section through appropriate space averaging.
Raman activities defined by Porezag and Pederson (1996) characterize phonon mode contributions to
the intensity of peaks in Raman spectra. These intensities depend on some other factors such as the
temperature and incident light wavelength. It is important to specify these parameters in order to
simulate a realistic Raman spectrum that can be compared to experimental results.
The Raman susceptibility tensor from which the Raman intensity can be calculated is defined as:
where:
χ(1) is the first order dielectric susceptibility
ν is a phonon eigenvector (the direction in which atoms, I, at equilibrium positions R, move under
excitation of a phonon mode, m in a unit cell with volume V)
Greek subscripts denote Cartesian directions
The derivative of χ(1) inside the sum can be calculated in CASTEP using a finite difference approach,
making small atomic displacements corresponding to the eigenvectors of each Raman active mode.
However, the derivative can be written in full as:
where:
ϵ is the electric field of a photon in a Raman experiment exciting the phonon
E is the total energy of the system.
We can use the (2n+1) theorem in quantum mechanics to evaluate this in a more direct manner, see
Miwa (2011). This theorem states that, under a perturbation, if the perturbed wavefunctions are known
to order n then we can evaluate the perturbed energy up to order 2n+1. In particular, in a CASTEP
phonon calculation, we need to evaluate the perturbed wavefunctions up to first order and therefore
we should have access to the 3rd order energy perturbation, E(3), from which we can compute the
susceptibility derivatives.
For Raman intensities, the full expression for this third derivative is given by equations (28)-(31) of Miwa
(2011). These expressions are implemented in CASTEP for norm-conserving pseudopotentials.
The finite difference method and the "2n+1" linear response method lead to the same Raman
susceptibility tensor, however the mathematical, and hence computational, routes are rather different.
In the finite difference method, the phonon eigenvectors are calculated during a phonon task, and are
then used to calculate the third order derivative for each mode one at a time. This means that there is a
separate Raman intensity calculation for each Raman active mode to obtain its intensity.
Computationally, this is usually the most time consuming part of a Raman intensity calculation.
Alternatively, in the linear response method, we first precompute the 9 second order derivatives with
respect to electric field: equation 19 of Miwa (2011) and then construct the full set of Raman tensors
during the phonon perturbations: equation (28) of Miwa (2011). Computationally, this means that in the
linear response Raman intensity method there is an initialization stage and each phonon perturbation
calculation has slightly more work to do than in the finite difference method. However, the
computational work is then complete, unlike in the finite difference method where additional electric
field perturbations are required for every Raman intensity.
For small cells the linear response method will be slower as the initialization stage and the extra cost of
each phonon perturbation does not outweigh the effort of calculating the finite difference Raman
tensor. For example, a linear response calculation for the primitive cell of boron nitride in the diamond
structure takes about twice as long as the finite displacement calculation. However, the linear response
method quickly becomes significantly more efficient as the number of degrees of freedom increases.
Calculations for molecular crystals containing 50-100 atoms per unit cell can be an order of magnitude
faster using the linear response method compared to the finite difference method.
Note: It is important to remember that Raman intensities are effectively third order derivatives, so to
obtain reasonable results very accurate calculations are required. Both the energy cutoff and k-point
sampling can contribute significantly to a convergence error, so a careful convergence test is
necessary to obtain reliable Raman intensities.
Thermodynamic calculations
The results of a calculation of phonon spectra can be used to compute energy (E), entropy (S), free
energy (F), and lattice heat capacity (Cv) as functions of temperature. The CASTEP total energy yields the
total electronic energy at 0 K. The vibrational contributions to the thermodynamic properties are
evaluated to compute E, S, F, and Cv at finite temperatures as discussed below.
Note: Thermodynamic calculations can be performed only if the system is in the ground state, that is,
geometry optimization is fully converged. This means that all the phonon eigenfrequencies must be
real and non-negative.
When you perform a vibrational analysis with CASTEP the results of the thermodynamic calculations can
be visualized using the thermodynamic analysis tools.
The formulas below are based on work by Baroni et al. (2001).
The temperature dependence of the energy is given by:
Eq. CASTEP 80
where Ezp if the zero point vibrational energy, k is the Boltzmann constant, ħ is the Planck constant and
F(ω) is the phonon density of states. Ezp can be evaluated as:
Eq. CASTEP 81
A popular representation of the experimental data on heat capacity is based on the comparison of the
actual heat capacity to that predicted by the Debye model. This leads to the concept of the temperature
dependent Debye temperature, ΘD(T). Heat capacity in Debye model is given by (Ashcroft and Mermin,
1976):
Eq. CASTEP 85
where N is the number of atoms per cell. Thus, the value of the Debye temperature, ΘD, at a given
temperature, T, is obtained by calculating the actual heat capacity, Eq. CASTEP 84, then inverting Eq.
CASTEP 85 to obtain ΘD.
accessible surface area) approximation calculates both of these terms, so they are proportional to the
surface area of the cavity.
Dziedzic et al. (2011) includes a brief description of the model, discussion of its implementation, and
comparison to other approaches. Dziedzic et al. (2013) provides a more detailed description of the
choice of numerical parameters. Fox et al. (2014) presents an example application to solvation of an
entire protein.
Setup tab
The Setup tab allows you to choose the type and quality of calculation that CASTEP performs, and other
basic input options such as exchange-correlation functional, spin polarization, and total charge.
Task: Select the type of calculation for CASTEP to perform. Available options are:
TS for GGA PBE, PBESOL, RPBE, BLYP, PW91; also for PBE0 and B3LYP
Grimme for GGA PBE, GGA BLYP, and B3LYP
OBS for GGA PW91 and LDA
D4 for GGA PW91, PBE, PBESOL, RPBE; also for PBE0, HSE03, HSE06, and RSCAN
MBD* for GGA PBE; also for HSE03, HSE06, and PBE0
The selected value might change when you select a different exchange-correlation functional; if there is
no support for the previous DFT-D scheme with the new functional. For example, if you selected the OBS
method and the LDA functional, the selection becomes invalid when you change the functional to GGA
PBE. In this case, the value of the dispersion correction method changes automatically to a valid option;
TS when possible, and D4 for the RSCAN functional. The DFT-D correction field becomes disabled if the
selected functional has no validated options, for example, for GGA WC. In this case, you can only apply
dispersion corrections by specifying custom values on the DFT-D tab of the CASTEP Electronic Options
dialog.
Note: The option selected automatically updates the Use custom DFT-D parameters option on the
DFT-D tab of the CASTEP Electronic Options dialog.
Spin polarization: Select how to treat spin density. Available options are:
Non-polarized also known as 'spin-restricted' calculation, uses the same orbitals for alpha and beta
spins.
Collinear also known as 'spin-unrestricted' calculation, separate calculation for majority and minority
spin projections.
Non-collinear the calculation allows spin density to vary the quantization axis in each point of space.
Use formal spin as initial: When selected, indicates that the calculation takes the initial value for the
number of unpaired electrons from the formal spin specified for each atom. The calculation optimizes
this starting value. If you do not select this option, you can specify the initial value for the number of
unpaired electrons. Default = checked.
Note: You can only specify Use formal spin as initial if you select Non-collinear or Collinear for the
Spin polarization.
Note: CASTEP only uses atom-resolved initial spin if the Electronic minimizer is Density mixing on the
SCF tab of the CASTEP Electronic Options dialog or if you select DFT+U.
Use DFT+U: When checked, CASTEP uses DFT+U (also known as the LDA+U formalism). You can specify
the actual values of U on the Hubbard U tab of the Electronic Configuration dialog.
Note: The DFT+U formalism is not compatible with NMR or polarizability calculations. You cannot use
this formalism with mixture atoms or real-space pseudopotentials.
Include spin-orbit coupling: When selected, adds spin-orbit coupling to the Hamiltonian.
Note: You can only specify Include spin-orbit coupling if you select Non-collinear or Collinear for the
Spin polarization. You can only use the spin-orbit coupling formalism for norm-conserving
pseudopotentials, either on-the-fly generated or tabulated, with no LDA+U corrections. In addition, it
is not compatible with calculations of:
NMR
Dipole corrections
Electron localization function
Linear response, including polarizability calculation
Electronic excitations
Partial density of states
Hirshfeld population analysis
You cannot visualize the electrostatic potential.
Metal: When selected, this indicates that the system is metallic so that you can create and edit empty
bands. Default = checked.
If you do not specify that a system is metallic:
Fix occupancy on the SCF tab of the CASTEP Electronic Options dialog is selected.
The default k-point separation value is coarser.
Note: In some calculations, it is not sufficient to use fixed orbital occupancies if DFT results indicate
that the system under investigation has a band gap. This applies to calculations that use the linear
response formalism for vibrational properties and to calculations of electric field responses
(polarizability and IR and Raman spectra). Attempting to study a metallic systems using fixed orbital
occupancies results in poor convergence and usually in unphysical results for IR and Raman
intensities and for polarizability. If this is the case, verify whether your system is metallic by running
either density of states or band structure calculations.
Charge: Specify the total charge on the unit cell.
Access methods
Access methods
Minimizer tab
The Minimizer tab allows you to specify the convergence level of the minimization and the external
pressure, and to control cell optimization.
Quality: Set the geometry optimization convergence thresholds for energy change, maximum force,
maximum stress, and maximum displacement between optimization cycles. The optimization will stop
when all of these criteria are satisfied.
Four sets of convergence thresholds are available:
Coarse
Medium
Fine
Ultra-fine
The values of the convergence thresholds in each set are given in the table below:
Alternatively, thresholds can be specified independently for Energy, Max. force, Max. stress, and Max.
displacement. If you enter your own values for any of these settings, the Quality is displayed as
Customized on both the Minimizer tab and the Setup tab of the CASTEP Calculation dialog.
Note: When the main quality is set to Express the geometry optimization quality is also set to
Express. In this case energy tolerance will be displayed in eV/cell rather than eV/atom, the default
value is 0.001 eV/cell. All other tolerances are ignored and geometry optimization runs stop when the
energy change is less than the tolerance regardless of the values of forces or displacements.
Energy: Specify the convergence threshold for the maximum energy change, in eV/atom, during the
geometry optimization. The units are eV/cell if the Quality is set to Express.
Max. force: Specify the convergence threshold for the maximum force, in eV/Å, during the geometry
optimization. If the Quality is set to Express, this tolerance by default will be set to a very high value and
will effectively be ignored in the calculation.
Max. stress: Specify the convergence threshold for the maximum stress, in GPa, during the geometry
optimization. If the Quality is set to Express, this tolerance by default will be set to a very high value and
will effectively be ignored in the calculation.
Max. displacement: Specify the convergence threshold for the maximum displacement, in Å, during the
geometry optimization. If the Quality is set to Express, this tolerance by default will be set to a very high
value and will effectively be ignored in the calculation.
Max. iterations: Specify the maximum number of geometry optimization cycles. If this number of cycles
is reached, then the calculation will stop, even if the convergence criteria are not satisfied.
Cell optimization: Specify whether the cell should be optimized and how. Default = None. Options are:
None - No cell optimization is requested and the Basis set for variable cell and Compressibility
options are not available.
Full - Complete optimization of the cell.
Fixed Volume - The cell is optimized while keeping the volume fixed, cell angles and cell lengths are
varied.
Fixed Shape - The cell is optimized while keeping the cell angles and cell length ratios fixed, the cell
volume is varied.
Basis set for variable cell: Specify the desired behavior of the basis set when the cell size changes.
Default = Fixed Basis Quality. Options are:
Fixed Basis Quality - The number of plane waves changes to maintain the fixed cutoff energy (finite
basis set correction is required)
Fixed Basis Size - The number of plane waves is kept constant, so the effective cutoff energy changes
(finite basis set correction is not required)
Tip: Fixed Basis Quality is the recommended option.
Compressibility: Specify an estimated compressibility value for the crystal under investigation in order
to accelerate the geometry optimization. Each entry in the dropdown list is interpreted as a preset
value:
Soft - 25 GPa, suitable for soft materials like molecular crystals
Medium - 200 GPa, suitable for most transition metals
Hard (default) - 500 GPa, suitable for minerals, ceramics, and hard metals
Note: CASTEP geometry optimization will respect constraints on the fractional coordinates of atoms.
These constraints may be imposed using the controls on the Atom tab of the Edit Constraints dialog.
Nonlinear constraints are respected only for P1 structures that do not contain mixture atoms and
only when the delocalized internals optimizer is used, that is, the Use delocalized internals checkbox
is checked on the Options tab of the CASTEP Geometry Optimization dialog.
Access methods
Options tab
The Options tab allows you to specify the minimization algorithm and the parameters controlling the
algorithm.
Algorithm: Specify the minimization algorithm. These algorithms are supported by CASTEP:
LBFGS
BFGS
Damped MD
TPSD
The LBFGS algorithm is generally recommended, especially for large systems. It implements a universal
sparse preconditioner that accelerates geometry optimization (Packwood et al., 2016). However, if there
are constraints TPSD might be a better choice (see the Geometry optimization theory topic for further
details).
The TPSD geometry optimization algorithm is a viable alternative for systems where the potential energy
surface deviates significantly from the quadratic form. TPSD offers robust convergence and is
particularly recommended when user-supplied constraints on lattice parameters are applied during cell
optimization (for example, when optimizing a solid-solid interface in a direction perpendicular to the
interface plane).
Note: The Damped MD algorithm is not used when cell optimization is performed.
Use line search: When checked, indicates that optimization should use the advanced line search
algorithm.
Note: This checkbox is checked by default, except when the main quality is set to Express. If you
specify the Express quality setting and then request line search, the Quality is displayed as
Customized on both the Minimizer tab and the Setup tab of the CASTEP Calculation dialog. In the
latter scenario, you must set convergence criteria for gradients and displacements to sensible values,
as the line search algorithm cannot function using only the energy change criterion.
Use delocalized internals: When checked, indicates that delocalized internal coordinates should be
used during geometry optimization, instead of Cartesian coordinates. This setting typically speeds up
the calculation by 2-3 times.
Note: The Use delocalized internals option is enabled only if the BFGS algorithm is selected.
Note: The delocalized internals minimizer does not currently support cell optimization. As a result,
CASTEP will automatically switch over to Cartesian coordinates when the cell is allowed to vary.
The delocalized internals minimizer respects nonlinear constraints only for P1 structures that do not
contain mixture atoms.
Damping algorithm: Specify the algorithm for the Damped MD method. The supported algorithms are:
Independent - the damping constant is chosen independently for each mode, so that all modes lose
energy at the same rate.
Coupled - the same damping constant is used for all modes at each step.
Steepest Descent
The Independent damping algorithm is recommended (see the Geometry optimization theory topic for
details).
Damping coefficient recalculation frequency: During the damped MD optimization, the damping
coefficients are recalculated occasionally to improve the damping rate. This value specifies the number
of MD steps between these recalculations.
Note: The Damping coefficient recalculation frequency value is not used by the steepest descent
algorithm.
Wavefunction extrapolation: Specify the method to be used for wavefunction extrapolation between
MD steps.
None - no extrapolation is performed.
1st VV - first-order extrapolation with variable coefficients.
2nd VV - second-order extrapolation with variable coefficients.
2/1 VV - alternate second- and first-order extrapolation with variable coefficients.
1st CC - first-order extrapolation with constant coefficients.
2nd CC - second-order extrapolation with constant coefficients.
2/1 CC - alternate second- and first-order extrapolation with constant coefficients.
Time step: Specify the time step to be used in the damped MD geometry optimization.
Auto-update time step: When checked, indicates that the time step will be automatically updated to
speed up the geometry optimization.
Note: The Time step and Auto-update time step controls are enabled only if the Damped MD
algorithm is selected.
Access methods
Stress tab
The Stress tab allows you to specify an external stress or hydrostatic pressure on the structure for
geometry optimization.
Note: The options on this tab are applicable only when cell optimization has been requested.
External stress: Specify any element of the external stress tensor on the structure, in GPa.
Equivalent hydrostatic pressure: Specify the hydrostatic pressure on the structure.
The hydrostatic pressure is one third of the trace of the stress tensor. When a value is specified here, the
diagonal elements of the stress tensor are modified so as to leave any anisotropic term unchanged.
Note: A positive value of external stress and hydrostatic pressure indicates that the structure is
compressed.
Reset external stress to zero: Sets all elements of the stress tensor to zero.
Access methods
Dynamics tab
The Dynamics tab allows you to specify the main parameters for controlling dynamics simulations.
Ensemble: Specify the ensemble to be used in the dynamics simulation.
NVE - dynamics at fixed volume and constant energy.
NVT - dynamics at fixed volume with a thermostat to maintain a constant temperature.
NPH - dynamics at constant enthalpy with a barostat to maintain a constant pressure.
NPT - dynamics with a thermostat to maintain a constant temperature and with a barostat to
maintain a constant pressure.
Initial velocities: Select the method of velocity initialization to be used for the dynamics calculation. The
available options are:
Random - set initial velocities to random value (default).
Current - set initial velocities to values currently defined on the structure.
Temperature: Specify the temperature of the dynamics simulation. For the NVE ensemble, the initial
random velocities of the atoms are scaled to this temperature.
Pressure: Specify the external pressure for the dynamics simulation in the NPH or NPT ensemble. For
Parrinello barostat full stress tensor can be specified on the Stress tab.
Time step: Specify the time step for the dynamics simulation.
Total simulation time: Specify the total time for the dynamics simulation. When modified the number
of steps is adjusted appropriately for the current time step.
Number of steps: Specify the total number of steps for the dynamics simulation.
Basis set for variable cell: Specify the desired behavior of the basis set when the cell size changes.
Fixed Basis Quality - the number of plane waves changes to maintain the fixed cutoff energy (finite
basis set correction is required).
Fixed Basis Size - the number of plane waves is kept constant, so the effective cutoff energy changes
(finite basis set correction is not required).
Tip: Fixed Basis Quality is the recommended option.
Access methods
Options tab
The Options tab allows you to specify parameters which control the details of the dynamics algorithm.
Thermostat: Specify the thermostat to be used for NVT or NPT dynamics. Available options are:
Langevin
Nose
NHL
Note: This option in enabled only if the NVT or NPT ensemble is selected on the Dynamics tab. The
Nosé-Hoover-Langevin (NHL) thermostat is only available for the NVT ensemble.
Barostat: Specify the barostat to be used for NPH or NPT dynamics. Available options are:
Andersen - scales the cell uniformly
Parrinello - allows variation in the cell shape
Note: This option in enabled only if the NPT or NPH ensemble is selected on the Dynamics tab.
Langevin constant: Specify the damping time of temperature fluctuations during Langevin NVT
dynamics. This time should normally be larger than the time scale of any motions of interest.
Nose Q ratio: Specify a scaling factor for the fictitious mass Q of the Nosé thermostat for Nosé or Nosé-
Hoover-Langevin NVT dynamics. A larger Q ratio implies increased damping of temperature fluctuations.
Cell time constant: Specify the characteristic cell time, which is used to estimate the mass associated
with the barostat. A larger value implies increased damping of pressure fluctuations.
Wavefunction extrapolation scheme: Specify the method to be used for wavefunction extrapolation
between molecular dynamics steps. This selection is not available when Use Extended Lagrangian (XL-
BOMD) is selected.
Stress tab
The Stress tab allows you to specify an external stress or hydrostatic pressure on the structure for
molecular dynamics.
Note: The options on this tab are available only when the NPT or NPH ensemble is selected on the
Dynamics tab.
External stress: Specify any element of the external stress tensor on the structure, in GPa.
Equivalent hydrostatic pressure: Specify the hydrostatic pressure on the structure.
The hydrostatic pressure is one third of the trace of the stress tensor. When a value is specified here, the
diagonal elements of the stress tensor are modified so that any anisotropic term remains unchanged.
Note: A positive value of external stress and hydrostatic pressure indicates that the structure is
compressed.
Note: When the Andersen barostat is selected on the Options tab only hydrostatic pressure is used
by CASTEP.
Reset external stress to zero: Sets all elements of the stress tensor to zero.
Access methods
means that the xx, yz (and zy) components of the strain tensor will be non-zero.
Note: This matrix is not editable.
Note: For trigonal symmetry, the orientation standard where C lies along the z-axis and A is in the xz
plane must be used. For further details on the lattice orientation standard and how to set it, see the
Viewing or changing the lattice parameters topic.
Access methods
Options tab
The Options tab allows you to specify the convergence level of the elastic constants calculation.
Quality: Sets the convergence thresholds for energy change, maximum force, and maximum
displacement between cycles. This is used for the geometry optimization runs that are required for the
elastic constants calculation. The calculation will stop when all of these criteria are satisfied.
Four sets of convergence thresholds are available; Coarse, Medium, Fine and Ultra-fine. The values of
each convergence threshold in each set are given in the table below:
Alternatively, thresholds can be specified independently. If user defined values are used, the Quality
displayed is Customized.
Note: When the main quality is set to Express the elastic constants quality is also set to Express. In
this case energy tolerance will be displayed in eV/cell rather than eV/atom, the default value is
0.001 eV/cell. All other tolerances are ignored and geometry optimization runs stop when the energy
change is less than the tolerance regardless of the values of forces or displacements.
Energy: Specify the convergence threshold for the maximum energy change, in eV/atom. The units are
eV/cell if the Quality is set to Express.
Max. force: Specify the convergence threshold for the maximum force, in eV/Å. If the Quality is set to
Express, this tolerance by default will be set to a very high value and will effectively be ignored in the
calculation.
Max. displacement: Specify the convergence threshold for the maximum displacement, in Å. If the
Quality is set to Express, this tolerance by default will be set to a very high value and will effectively be
ignored in the calculation.
Max. iterations: Specify the maximum number of elastic constants cycles. If this number of cycles is
reached, then the calculation will stop even if the convergence criteria are not satisfied.
Access methods
Note: When a transition state search is performed on a periodic system the unit cell is fixed.
Search protocol: Select the type of synchronous transit that will be performed from the dropdown list.
Available options are:
LST Maximum - performs a single linear synchronous transit (LST) maximization, bracketing the
maximum between the reactants and product
Halgren-Lipscomb - performs an LST maximization, followed by a single-line search minimization
LST/Optimization - performs an LST maximization, followed by a full conjugate gradient minimization
Complete LST/QST - performs an LST, followed by repeated conjugate gradient minimizations and
quadratic synchronous transit (QST) maximizations until a transition state has been located
QST/Optimization - starting from a QST, performs repeated maximizations and conjugate gradient
minimizations until a transition state has been located
Quality: Set the TS search convergence thresholds for the root mean square (RMS) forces on the atoms.
The search will stop when this criterion is satisfied.
Five quality settings are available:
Express
Coarse
Medium
Fine
Ultra-fine
The value of the convergence threshold for each setting is given in the table below:
RMS convergence: Specify the value, in eV Å-1, at which convergence is considered to take place in terms
of the RMS of the gradients.
Max. number QST steps: Specify the maximum number of QST maximization cycles. If this number of
cycles is reached, then the calculation will stop, even if the convergence criteria are not satisfied.
Optimize reactants and products: Specifies that the input structure for the reactant and product in a TS
search calculation should be optimized to a local minimum. Selecting this allows the TS search to start
from a "raw" trajectory, where the reactant and product structures have not already been optimized.
Note: This choice may lead to reactant and product structures deviating substantially from the input
structures, so it is recommended that you carefully check the final path.
Calculate partition functions: Specifies whether to calculate partition functions based on the Gamma-
point dynamical matrix for the reactants, products, and the transition state. This data will be stored in
the output 3D Atomistic Collection document.
Note: This is not available unless Optimize reactants and products is checked.
Help: Displays the Help topic in a browser.
Access methods
Alternatively, thresholds can be specified independently for Energy, Max. force, and Max.
displacement. If you enter your own values for any of these settings, the Quality is displayed as
Customized on both the CASTEP TS Confirmation dialog and the Setup tab of the CASTEP Calculation
dialog.
Note: When the main quality is set to Express the transition state confirmation quality is also set to
Express. In this case the energy tolerance will be displayed in eV/cell rather than eV/atom, the default
value is 0.002 eV/cell.
Energy: Specify the convergence threshold for the maximum energy change, in eV/atom, during the
geometry optimization. The units are eV/cell if the Quality is set to Express.
Max. force: Specify the convergence threshold for the maximum force, in eV/Å, during the geometry
optimization.
Max. displacement: Specify the convergence threshold for the maximum displacement, in Å, during the
geometry optimization.
Altering the value of any threshold is allowed and results in the Quality being set to Customized.
Max. images: The number of intermediate NEB images used during the transition state confirmation.
Help: Displays the Help topic in a browser.
Access methods
Electronic tab
The Electronic tab allows you to specify the parameters associated with the electronic Hamiltonian.
These include the basis set, pseudopotentials, SCF convergence criteria, and k-point set.
Energy cutoff: Select the precision used for the plane wave basis set.
Coarse
Medium
Fine
Ultra-fine
The actual value of the energy cutoff corresponding to the selected quality for the current structure also
displays.
Note: Calculation of the NMR parameters may require a higher Energy cutoff than for total energy or
geometry optimization calculations. Study the convergence of the results as you increase the Energy
cutoff value. Generally, a cutoff between 270 and 400 eV produces convergence of the shielding
constant to about 1 ppm. In some cases, you might need to increase the cutoff up to 550 eV.
SCF tolerance: Select the threshold used to determine whether an SCF has converged. Available options
and associated convergence thresholds are:
Coarse - 10-5 eV/atom
Medium - 2 × 10-6 eV/atom
Fine - 10-6 eV/atom
Ultra-fine - 5 × 10-7 eV/atom
Note: When the main quality is Express, the SCF tolerance is automatically Express. This corresponds
to 10-4 eV/cell.
Energy tolerances per: Specifies whether to calculate energy tolerances as the energy change per atom
(Atom is the default) or per unit cell. The Cell value is enforced when the main quality, and hence the SCF
tolerance quality, is Express.
k-point set: Define the number of integration points to use to integrate the wavefunction in reciprocal
space. Available options are:
Gamma - a single point at (0, 0, 0)
Coarse
Medium
Fine
Pseudopotentials: Select the type of pseudopotential to use. Two different types of pseudopotentials
are available, ultrasoft and norm-conserving; both of them are available as either generated on the fly
(OTFG) potentials or as tabulated potentials. OTFG potentials are generally preferred for reasons of
accuracy and consistency. NMR calculations required OTFG potentials.
OTFG ultrasoft (default) - the preferred option, ultrasoft pseudopotentials allow calculations with
lower energy cutoffs. On-the-fly generation produces potentials that are consistent between solid
state and pseudo-atom calculations since they use the same exchange-correlation functional
throughout. The OTFG ultrasoft pseudopotentials introduced in Materials Studio 8.0 provides results
in close agreement with all electron calculations and is very accurate for describing ground state
structures.
OTFG norm conserving - typically harder and therefore more computationally expensive than
ultrasoft pseudopotentials. Linear response-based calculations of phonon properties and of
polarizabilities and calculations with nonlocal exchange-correlation functionals (for example,
screened exchange) require norm-conserving potentials.
Ultrasoft - tabulated ultrasoft pseudopotentials, provided for compatibility with earlier versions of
Materials Studio.
Norm conserving - tabulated norm-conserving pseudopotentials, provided for compatibility with
earlier versions of Materials Studio.
Notes:
Do not use the PW91 exchange-correlation functional when requesting on the fly generation of
pseudopotentials. When you require a GGA functional, use PBE, RPBE, WC, BLYP, or PBESOL.
When you specify the type of pseudopotential to use in this way, the calculation uses the default
pseudopotential of that type for each element.
Tip: Use the Electronic Options dialog to exercise more control over which pseudopotentials to use
for each element.
Relativistic treatment: The way OTFG pseudopotentials incorporate relativistic effects. Options are:
Schroedinger
Koelling-Harmon (default)
ZORA
Dirac
Note: Use the Dirac solver when requesting spin-orbit coupling and on the fly generated
pseudopotentials.
Use core hole: Specify whether to include core holes in the calculation. You can use core holes in core
level spectroscopy calculations. Define core holes using the Core Hole tab of the Electronic Configuration
dialog.
Note: For calculations exploiting core holes, you must use on-the-fly pseudopotentials.
If you do not select the Use core hole checkbox, the calculation does not preserve any core holes in
the input system.
Note: CASTEP calculations with core holes assume by default that an electron is removed from the
system, so it is treated as having a charge of 1 (or one more than specified in the Charge field on the
Setup tab of the CASTEP Calculation dialog). In some circumstances, for example when studying
impurities, it might be more appropriate to consider the system as charge neutral. In order to achieve
this you should specify the Charge of -1 on the Setup tab, in this way the CASTEP input file will have a
zero charge. It is also possible to edit the seedname.param file to specify a fractional charge. The
strength of the core hole can be adjusted by editing the seedname.cell file and modifying the
pseudopotential definition string (for example, change {1s1} to {1s0.5} to use a core hole with
the charge of 0.5).
Use solvation model: When selected, uses the implicit solvation model. The Solvent tab of the
CASTEP Electronic Options dialog allows you to define the solvent characteristics.
Note: The implicit solvation model allows you to simulate a solvent environment only for "molecule
in a box" calculations. Materials Studio automatically positions isolated molecule structures roughly
in the center of a cell. Ensure that the cell is large enough to contain a significant amount of vacuum
surrounding the molecule.
Note: Use at least a Fine FFT grid Density for solvation calculations; the Precise density is even better.
You can specify the FFT grid Density on the Basis tab of the CASTEP Electronic Options dialog. Also use
an increased value of Augmentation density scaling factor, for example 1.5-2.0 instead of the default
value of 1.0.
An increased FFT density and energy cutoff can help to eliminate convergence errors for the multigrid
solver for electrostatics in a solvent.
More...: Provides access to the CASTEP Electronic Options dialog, where you can configure further
details of the electronic Hamiltonian.
Access methods
Basis : Provides access to the plane-wave energy cutoff, FFT grid specification, and finite basis set
correction.
SCF: Provides access to the parameters that control the electronic minimization algorithm.
k-points: Provides more detailed access to the k-point set.
Potentials: Provides control over which pseudopotentials to use for each element in the structure.
DFT-D: Provides access to the parameters for van der Waals dispersion correction calculations.
Solvent: Provides access to the parameters for solvation calculations.
Help: Displays the Help topic for the current tab.
Access methods
Basis tab
The Basis tab provides access to basis set information, including the plane-wave energy cutoff, FFT grid
specification, and finite basis set correction.
Use custom energy cutoff: When selected, you can specify a custom value for the energy cutoff in the
adjacent text box. Otherwise the text box displays the energy cutoff appropriate for the Energy cutoff
quality parameter on selected on the Electronic tab.
Density: Specify the quality of the FFT grid. Three options are available:
Standard
Fine
Precise
The FFT grid parameters are chosen such that they provide a good representation of the electron
density. The Precise level provides an exact representation of the density for the given energy-cutoff.
The Standard and Fine levels use a reduced FFT grid to speed up the calculation. Some aliasing occurs as
a result, but the effect on the SCF solution is usually small. The FFT grid parameters used in the CASTEP
calculation display as derived, using the current energy-cutoff and unit cell parameters.
Augmentation density scaling factor: Specify a finer grid for representation of the augmentation charge
density in the atomic core region. By default, CASTEP uses the same grid as for the standard electron
density. It is reasonable in some cases to increase the augmentation grid by a factor of up to 4, to
improve the accuracy of calculations. You can control this with the Quality option on the Setup tab of
the CASTEP Calculation dialog.
Apply finite basis set correction: Specify when to use a finite basis set correction in the evaluation of
energy and stress. You can control this with the Quality option on the Setup tab of the CASTEP
Calculation dialog.
This correction gives a volume dependent addition to the energy that may result in a significant
correction to any calculated stress. It is particularly important where you require an accurate stress or
cell geometry optimization. Three options are available:
Smart - apply the correction only when stress is required. This includes any cell geometry
optimization.
Always - apply the correction to all calculations.
Never - do not apply a correction, even when calculating stress. Any result for stress or cell geometry
may be unreliable.
Correction mode: Specify how to calculate the finite basis set correction. The evaluation of the
correction requires an estimate for the derivative of the total energy with respect to the logarithm of the
plane-wave energy cutoff. The options are:
Automatic - automatically calculates the energy derivative using numerical differentiation.
Manual - allows you to manually specify the energy derivative.
Numerical differentiation using: Specify the number of energy points to use in evaluating the energy
derivative required for the finite basis set correction. This evaluates a converged SCF energy at each
point.
Note: Available only if you select the Automatic Correction mode.
Energy derivative w.r.t. cutoff: Specify the value of the energy derivative required for the finite basis set
correction.
Note: Available only if you select the Manual Correction mode.
Access methods
SCF tab
The SCF tab on the CASTEP Electronic Options dialog provides access to parameters that control the
electronic minimization algorithm.
SCF tolerance: Specify the SCF convergence threshold. This overrides the SCF tolerance option specified
on the Electronic tab.
Note: When calculating linear response properties (for example phonons, polarizabilities, NMR
properties), it is important that the SCF convergence threshold is sufficiently strict from the
computational point of view to accurately calculate these properties. When you request the
calculation of linear response properties, CASTEP uses a convergence threshold of 10-8 eV atom-1,
unless the specified threshold is lower. To force CASTEP to use a less stringent threshold, manually
add the ELEC_ENERGY_TOL parameter to the input parameters (.param) file for the property
calculation.
Max. SCF cycles: Set the maximum number of SCF iterations allowed for an energy calculation. The
energy calculation ends after the specified number of iterations, even if the SCF has not converged.
Convergence window: The convergence tolerance window for SCF. This is the number of SCF iterations
in a row for which convergence criteria are fulfilled. This parameter prevents occasional oscillations.
Electronic minimizer: Specify the electronic minimization method to use for the SCF calculation. Two
options are available, Density Mixing and All Bands/EDFT.
Tip: The Density Mixing option is recommended, particularly for variable occupancy calculations. The
All Bands/EDFT method is usually slower and requires more memory. However, use the EDFT scheme
for metallic systems where converging electronic structure using the Density Mixing option proves
problematic.
Charge: Specify the amplitude for charge density mixing (that is, the amplitude of the output charge
density to mix with the input charge density to obtain input for the next iteration).
Note: Available only when you select the Density Mixing minimization method.
Spin: Specify the amplitude for spin density mixing, that is, the amplitude of the output spin density to
mix with the input spin density to obtain input for the next iteration.
Note: Available only for spin-polarized calculations using the Density Mixing minimization method.
More...: Provides access to the CASTEP Density Mixing Options dialog, where more detailed control over
the Density Mixing electronic minimization method is available.
Note: Available only when you select the Density Mixing minimization method.
Fix occupancy: When selected, this fixes thee electronic occupation numbers during electronic
minimization. CASTEP includes only the lowest occupied states in the calculation. Default = checked.
More...: Provides access to the CASTEP Occupancy Options dialog, where more detailed control over the
occupancies is available.
Apply dipole corrections: Select the type of dipole corrections to apply from the list. Available options
are:
None (default)
Non self-consistent
Self-consistent
Self-consistent is the most accurate option, as this affects the electrostatic potential as well as the total
energy and its gradients. This is the only option that is suitable to investigate the effect of dipole
corrections, for example on surface workfunctions.
Tip: It is recommended to use the All Bands/EDFT electronic minimization scheme when applying a
self-consistent dipole correction to an elongated cell of a slab representing a metal surface; the
Density mixing minimization scheme may fail to converge for such systems.
Use Non self-consistent for cases where it is hard to converge the SCF with Apply dipole corrections. You
can only use this option for Energy calculations, it is not compatible with phonon calculations.
Note: CASTEP applies dipole corrections only if the system is either molecule in the box or a periodic
slab with the vacuum layer of at least 8 Å.
Note: To apply dipole corrections, CASTEP converts the symmetry of the system to P1 to bring the
molecule (or slab) into the center of the unit cell.
Access methods
Note: The spin will be fixed throughout the calculation (including possible steps of geometry
optimization) if the checkbox is unchecked.
Tip: In spin-polarized systems there may be more than one stable SCF solution with different total
spin values. By setting the initial spin on the Setup tab and by fixing the total spin value for a number
of SCF iterations, it is possible to exert some control over which solution is found.
Help: Displays the Help topic in a browser.
Access methods
k-points tab
The k-points tab provides access to more detailed options for controlling the k-point set used in CASTEP
calculations.
You can specify the Monkhorst-Pack k-point grid to use in several ways. For cubic cells and for the C
direction of hexagonal cells, the even and odd grids for the Monkhorst-Pack scheme give the same
number of k-points. However, the even grid provides better sampling and CASTEP always uses this
automatically under these conditions. This achieves a good grid with k-point separation at (or less than)
the specified target more economically. As a consequence, such lattices may have much finer
separations than requested as CASTEP excludes odd grids - even if they would have been closer to the
specified separation - and takes the better even grids in preference.
Gamma point only: When selected, indicates that the density of states calculation uses a single k-point
at (0,0,0).
Quality: When selected, generates the k-point grid using a k-point separation appropriate to the
specified quality level. Select the desired quality level from the list. Available options are:
Coarse
Medium
Fine
The k-point separations associated with the three Quality settings depend on whether you select the
Metal checkbox on the Setup tab:
Separation: When selected, generates the k-point grid according to the specified k-point separation.
Specify the k-point separation, in Å-1, in the associated text box.
Note: When you select the Separation, the Monkhorst-Pack parameters are derived to give the
specified separation between neighboring grid points.
Custom grid parameters: When selected, generates the k-point grid using the Monkhorst-Pack grid
parameters and the origin shift in fractional reciprocal space coordinates specified in the Grid
parameters and Origin shift text boxes, respectively.
Grid parameters: Specify the Monkhorst-Pack grid parameters in each of the lattice directions.
Actual spacing: Displays the k-point separation, in Å-1, resulting from the currently specified
Monkhorst-Pack grid parameters in each of the lattice directions.
Origin shift: Specify the offset of the Monkhorst-Pack grid in fractional reciprocal space coordinates.
Note: The Grid parameters and Origin shift controls are enabled only if you select the Custom grid
parameters option.
Display points...: Displays the number and fractional coordinates of the reciprocal space mesh points
generated using the specified parameters.
Note: The actual set of k-points used in the calculations may change if the symmetry of the system
changes or if you request a DFT+U calculation.
Access methods
Potentials tab
The Potentials tab allows you to control which pseudopotentials to use for each element in the
structure.
Scheme: Select the type of pseudopotential to use. Two different types of pseudopotentials are
available, ultrasoft and norm-conserving. Both of these are available as either generated on the fly
(OTFG) potentials or as tabulated potentials. OTFG potentials generally provide better accuracy and
consistency. You can only perform NMR calculations with OTFG potentials.
OTFG ultrasoft (default) - the preferred option, ultrasoft pseudopotentials allow you to perform
calculations with lower energy cutoffs. On-the-fly generation produces potentials that are consistent
between solid state and pseudo-atom calculations since they use the same exchange-correlation
functional throughout. The set of OTFG ultrasoft pseudopotentials introduced in Materials Studio
8.0 provides results in close agreement with all electron calculations and is very accurate for
describing ground state structures.
OTFG norm conserving - typically harder and therefore more computationally expensive than
ultrasoft pseudopotentials. Norm-conserving potentials are required for linear response-based
calculations of phonon properties and of polarizabilities and for calculations with nonlocal exchange-
correlation functionals (for example, screened exchange).
Ultrasoft - tabulated ultrasoft pseudopotentials, provided for compatibility with earlier versions of
Materials Studio.
Norm conserving - tabulated norm-conserving pseudopotentials, provided for compatibility with
earlier versions of Materials Studio.
Note: The latest set of OTFG settings for ultrasoft pseudopotentials was developed to minimize the
error with respect to fully converged all-electron DFT calculations. The error achieved by this set is
0.4 meV/atom, which puts CASTEP among the most accurate pseudopotential codes available. A full
definition of the test framework and the meaning of the error is given by Lejaeghere et al. (2014), and
on the website of the Delta project: http://molmod.ugent.be/deltacodesdft. The files that
correspond to this set have the _2017R2 suffix. This is the recommended set of OTFG ultrasoft
potentials. Use files with the _2017R2ncp suffix when you require norm-conserving potentials. This
set has a 1.1 meV/atom error in the Delta project tests. In addition, provides a set of "high
throughput" ultrasoft OTFG settings for scripting. This QC5 set generally requires lower cutoff energy
and offers faster but less accurate calculations (1.9 meV/atom error in the Delta project tests).
Note: Do not use the PW91 exchange-correlation functional when requesting on the fly generation of
pseudopotentials. PBE, RPBE, WC, BLYP, or PBESOL are better options when you require a GGA
functional.
Note: When you specify the type of pseudopotential to use in this way, the default pseudopotential
of that type for each element will be used in the calculation.
Representation: Choose between Reciprocal space and Real space representations of the
pseudopotentials.
Tip: The real space representation may give a performance advantage for large systems.
More...: Provides access to the CASTEP Real Space Potentials dialog, which allows you to specify details
of the transformation of pseudopotentials to a real space representation.
Note: This is enabled only if you select Real space from the Representation list.
Selected potentials: Displays information about the selected pseudopotentials for each element in a
grid.
There is a row in the grid for each element displaying the element type, current pseudopotential,
energy, and valence (Z). To select a different pseudopotential for a particular element, click in the
appropriate pseudopotential grid cell and choose a new pseudopotential from the list.
If you change the default pseudopotentials, the Scheme displays as Custom.
Show potentials for all elements: When selected, the Selected potentials list displays information
about all elements in the periodic table for which pseudopotentials are available. Otherwise, only
elements present in the current structure are displayed in the list. Default = unchecked.
View: Displays a modal view on the selected pseudopotential file.
External Potential
Electric field: Specify the external electric field strength, in eV/Å/e.
Field direction (X Y Z): Vector specifying the direction of the electric field.
Note: Only systems that contain molecules or slabs are appropriate for electric field calculations.
For slabs, the field must be along the surface normal. In both cases, you must position atoms in the
middle of the cell and the symmetry must be P1.
Note: The field direction is specified in Cartesian coordinates. The values of the X, Y, and Z direction
components of the electric field are interrelated. The three field components define a unit vector.
When you start a job, the direction components are normalized to give a unit vector, which is then
multiplied by the field magnitude to give the actual field components. Entering a vector with all zero
components has the same effect as entering a (111) vector.
Access methods
Note: The real space core radius determines the radius of real space integration to be used in the
calculation. It should be larger than the core radius used in pseudopotential generation, normally by
a factor of 1.5 to 2.0, so as to overcome the aliasing problems inherent in the real space integration
method.
Reset all core radii to defaults: Sets all core radii to the default values.
Use custom energy cutoff for transformation: When checked, a value may be specified for the energy
cutoff used in the transformation to the real space form of pseudopotentials. Otherwise a value will be
calculated automatically based on the plane-wave energy cutoff to be used in the calculation. Default =
unchecked.
Impose tolerance: When checked, the real space pseudopotential transformation will be tested to
ensure it is of the quality specified in the dropdown list. Default = checked.
If the transformation is not of the requested quality, then the CASTEP calculation will not proceed. This
should ensure that the results obtained from a real space pseudopotential calculation accurately
reproduce those of a similar calculation using reciprocal space pseudopotentials.
Note: If this option in unchecked, the CASTEP calculation will be performed regardless of the quality
of the real space pseudopotential transformation.
Help: Displays the Help topic in a browser.
Access methods
DFT-D tab
The DFT-D tab allows you to prepare customized parameters for van der Waals dispersion corrections.
This can involve the definition of DFT-D corrections for exchange functionals that Materials Studio does
not usually support, the definition of support for additional elements, or changing existing parameters.
Use custom DFT-D parameters: Specify whether to use custom DFT-D parameters. When selected, you
can choose the van der Waals scheme to use from the list, options are:
TS (default)
Grimme
OBS
MBD*
Note: You cannot customize the parameters of the D4 method.
Note: The option selected automatically updates the Use method for DFT-D correction option on the
Setup tab of the CASTEP Calculation dialog.
Note: The schemes that rely on calculated electron densities, namely TS and MBD*, may not work
well in calculations with small unit cells, where the diameter of van der Waals spheres is comparable
to the lattice parameters. The MBD* scheme is particularly sensitive to such situations. In such cases,
the CASTEP output file contains a warning explaining the problem. The MBD* scheme can also fail if
the geometry is far from equilibrium, in which case it is useful to perform a preliminary geometry
optimization with the TS scheme. The main area of application of the MBD* scheme is molecular
crystals, so it may experience problems with highly ionic systems.
Atomic parameters
This table allows you to edit the dispersion correction parameters for each of the atomic species. The
type and number of parameters depend on the DFT-D scheme selected. Refer to the original literature
for the meaning and usage of each parameter.
The units for all parameters are in eV for energies and in Å for lengths. The available columns are:
Atom: Denotes the element to edit.
C6 (eV Å6): Available for TS, MBD*, and Grimme schemes.
R0 (Å): Available for TS, MBD*, and Grimme schemes.
alpha (Å3): Available for TS, MBD*, and OBS schemes.
I (eV): Available for the OBS scheme.
Rvdw (Å): Available for the OBS scheme.
Note: The Atomic parameters table only lists the elements in the active document.
Note: Each scheme includes a radius (R0 for TS and Grimme, Rvdw for OBS) that must be non-zero for
each of the active elements.
Scheme parameters
Scheme parameters are dimensionless numbers that define the functional form of each scheme.
Refer to the original literature for the meaning and usage of each parameter.
Depending on the van der Waals scheme selected for the custom DFT-D parameters setting, the Scheme
parameters that you can specify are:
TS: sR, d
Grimme: s6, d
OBS: lambda, n
MBD*: sR
Note: The default values for these parameters depend on the exchange correlation functional
selected on the Setup tab of the CASTEP Calculation dialog.
Note: Both scheme parameters must be non-zero for the calculations to start and for Materials
Studio to write to the input files.
Reset All: Restores the default settings for each element in the grid.
Access methods
Solvent tab
The Solvent tab allows you to configure a simulated solvent environment for the calculation.
Solvent: Select a solvent to use as the solvent environment for the calculation. Available options are:
Acetone
Acetonitrile
Benzene
Carbon tetrachloride
Chlorobenzene
Chloroform
Cyclohexane
Diethyl ether
Dimethyl sulfoxide
Ethanol
Methanol
Methylene chloride
n-hexane
n-hexadecane
Nitrobenzene
Pyridine
Water
When you select a solvent, the Dielectric constant field automatically updates with the appropriate
value.
Note: If you specify the Dielectric constant parameter to a value different from that dictated by the
selected solvent, the Solvent displays as Customized.
Dielectric constant: Specify a value for the solvent dielectric constant. This parameter automatically
updates to the appropriate value when you select a solvent from the Solvent dropdown list. However,
you can enter your own custom value if you want.
Include apolar term: Request calculation of apolar (non-electrostatic) contributions to the solvation
energy. These contributions arise from cavitation and solute-solvent dispersion-repulsion. Calculations
without the apolar term cannot reliably estimate solvation energies, especially for neutral molecules,
and particularly so if these molecules are large.
Note: The Solvent surface tension and Account for solute-solvent dispersion repulsion controls are
enabled only when you select the Include apolar term checkbox.
Solvent surface tension: The surface tension of the solvent (in N/m) for evaluation of the apolar term.
Account for solute-solvent dispersion repulsion: The CASTEP solvation model suggests scaling of the
surface tension that enters the expression for the apolar term, to account for the solute-solvent
dispersion-repulsion. The default dimensionless scaling factor is 0.281075 (see Dziedzic, 2011).
Note: Use at least a Fine FFT grid Density for solvation calculations; the Precise density is even better.
You can specify the FFT grid Density on the Basis tab of the CASTEP Electronic Options dialog. Also use
an increased value of Augmentation density scaling factor, for example 1.5-2.0 instead of the default
value of 1.0.
An increased FFT density and energy cutoff can help to eliminate convergence errors for the multigrid
solver for electrostatics in a solvent.
Access methods
Properties tab
The Properties tab allows you to select the properties that will be computed as part of a CASTEP
calculation.
Choose the properties you wish to compute by checking the appropriate checkboxes in the list. Once
you select a properties checkbox, additional options may be displayed below the list.
Access methods
Separation: Specify the approximate separation between k-points in Å-1. The default depends on the
selected k-point set.
More...: Provides access to the CASTEP Band Structure Options dialog, which allows you to set
additional parameters, including the reciprocal space path.
Access methods
Note: There is an important difference between standard DFT calculations with local exchange-
correlation potentials and the nonlocal exchange case. The potential used in the latter scenario
depends on the SCF k-points, while in the former case, the potential depends only on electron
density. This difference can make all properties calculations a lot more expensive in terms of memory
usage and CPU time, so it is advisable to limit the number of SCF k-points that you use in such
situations. This is particularly relevant for small unit cells, where the default settings may generate a
very large k-point set.
Functional: Select the type of exchange-correlation potential to be used in the band structure
calculation.
Choose the type of functional from the first dropdown list. Available options are:
LDA
GGA
HF
HF-LDA
sX
sX-LDA
PBE0
B3LYP
HSE03
HSE06
Then, if appropriate, select the specific functional from the second dropdown list. Depending on the
functional type selected from the first dropdown list, some of the following options may be available:
CA-PZ
PBE
RPBE
PW91
WC
PBESOL
BLYP
Note: This option is enabled only when Use separate XC functional for band structure calculation is
checked.
Band energy tolerance: Specify the convergence criterion for electronic eigenvalues during the band
structure calculation. The calculation will stop when all electronic eigenvalues change by less than this
value over a single iteration of the electronic structure.
Note: More information about the symmetry labels used in the Band Structure Options dialog can be
found in Bradley and Cracknell (1972).
Path...: Provides access to the Brillouin Zone Path dialog, which allows you to specify a custom
reciprocal space path.
Help: Displays the Help topic in a browser.
Access methods
More...: Provides access to the CASTEP Core Level Spectroscopy Options dialog, which allows you to set
additional parameters, including the k-point set specification.
Access methods
Separation: When selected, generates the k-point grid according to the specified k-point separation.
Specify the k-point separation, in Å-1, in the associated text box.
Note: When you select the Separation, the Monkhorst-Pack parameters are derived to give the
specified separation between neighboring grid points.
Custom grid parameters: When selected, generates the k-point grid using the Monkhorst-Pack grid
parameters and the origin shift in fractional reciprocal space coordinates specified in the Grid
parameters and Origin shift text boxes, respectively.
Grid parameters: Specify the Monkhorst-Pack grid parameters in each of the lattice directions.
Actual spacing: Displays the k-point separation, in Å-1, resulting from the currently specified
Monkhorst-Pack grid parameters in each of the lattice directions.
Origin shift: Specify the offset of the Monkhorst-Pack grid in fractional reciprocal space coordinates.
Note: The Grid parameters and Origin shift controls are enabled only if you select the Custom grid
parameters option.
Display points...: Displays the number and fractional coordinates of the reciprocal space mesh points
that would be generated using the currently specified parameters.
Note: The actual set of k-points that will be used in the calculations may be altered if the symmetry of
the system changes.
Band energy tolerance: Specify the convergence criterion in eV for electronic eigenvalues during the
core level spectroscopy calculation. The calculation will stop when all electronic eigenvalues change by
less than this value over a single iteration of the electronic structure.
Help: Displays the Help topic in a browser.
Access methods
Quality: When selected, generates the k-point grid using a k-point separation appropriate to the
specified quality level. Select the desired quality level from the list. Available options are:
Coarse
Medium
Fine
The k-point separations associated with the three Quality settings are as follows:
Separation: When selected, generates the k-point grid according to the specified k-point separation.
Specify the k-point separation, in Å-1, in the associated text box.
Note: When you select the Separation, the Monkhorst-Pack parameters are derived to give the
specified separation between neighboring grid points.
Custom grid parameters: When selected, generates the k-point grid using the Monkhorst-Pack grid
parameters and the origin shift in fractional reciprocal space coordinates specified in the Grid
parameters and Origin shift text boxes, respectively.
Grid parameters: Specify the Monkhorst-Pack grid parameters in each of the lattice directions.
Actual spacing: Displays the k-point separation, in Å-1, resulting from the currently specified
Monkhorst-Pack grid parameters in each of the lattice directions.
Origin shift: Specify the offset of the Monkhorst-Pack grid in fractional reciprocal space coordinates.
Note: The Grid parameters and Origin shift controls are enabled only if you select the Custom grid
parameters option.
Display points...: Displays the number and fractional coordinates of the reciprocal space mesh points
that would be generated using the currently specified parameters.
Note: The actual set of k-points that will be used in the calculations may be altered if the symmetry of
the system changes.
Use separate XC functional for DOS calculation: When checked, indicates that a different exchange-
correlation functional will be used for the density of states calculation from the one used in the self-
consistent calculation.
Tip: This feature is particularly useful for carrying out expensive screened exchange calculations:
perform the main run with the LDA or GGA functional, then invoke screened or exact exchange in non-
self-consistent mode during the density of states calculation.
Note: There is an important difference between standard DFT calculations with local exchange-
correlation potentials and the non-local exchange case. The potential used in the latter scenario
depends on the SCF k-points, while in the former case, the potential depends only on electron
density. This difference can make all properties calculations a lot more expensive in terms of memory
usage and CPU time, so it is advisable to limit the number of SCF k-points that you use in such
situations. This is particularly relevant for small unit cells, where the default settings may generate a
very large k-point set.
Functional: Select the type of DFT exchange-correlation potential to be used in the density of states
calculation.
Choose the type of functional from the first dropdown list. Available options are:
LDA
GGA
HF
HF-LDA
sX
sX-LDA
PBE0
B3LYP
HSE03
HSE06
Then, if appropriate, select the specific functional from the second dropdown list. Depending on the
functional type selected from the first dropdown list, some of the following options may be available:
CA-PZ
PBE
RPBE
PW91
WC
PBESOL
BLYP
Note: The Functional control is enabled only if the Use separate XC functional for DOS calculation
checkbox is checked.
Band energy tolerance: Specify the convergence criterion, in eV, for electronic eigenvalues during the
density of states calculation. The calculation will stop when all electronic eigenvalues change by less than
this value over a single iteration of the electronic structure.
Help: Displays the Help topic in a browser.
Access methods
Tip: When only one set with the correct name is present in the structure, Materials Studio assumes
that the rest of the atoms should be treated as a second set. Thus, in the case of a molecule-surface
interaction, it is sufficient to create a single set that comprises all the atoms in the molecule.
Atomic densities: When selected, indicates that only the density difference with respect to the sum of
atomic densities will be calculated.
Sets of atoms: When selected, indicates that only the density difference with respect to the sum of the
densities of sets of atoms will be calculated.
Both atomic densities and sets of atoms: When selected, indicates that both kinds of density
difference will be calculated.
Access methods
NMR selection
Checking the NMR checkbox on the Properties tab displays options for calculating shielding tensors and
electric field gradients.
Note: NMR in CASTEP is part of the separately licensed module NMR CASTEP. NMR calculations can
only be performed if you have purchased this module.
Shielding: Determines the chemical shielding tensor for each atom in the system.
Note: The shielding tensor is ill-defined for spin-polarized calculations. NMR calculations are allowed
wth a spin-polarized setting for a paramagnetic system, but you should check that the final electronic
state is indeed paramagnetic. Similarly, such calculations are not recommended for metallic systems
without a band gap; however, it is legitimate to use variable orbital occupancies and metallic settings
while modeling an insulator.
EFG: Determines the electric field gradients for each atom in the system.
J-coupling: Determines the J-coupling between the selected atom(s) and all other atoms in the system.
At least one atom must be selected in the input 3D Atomistic document.
Note: J-coupling can be calculated only for non-magnetic systems. The Relativistic treatment of the
OTFG pseudopotentials (on the Electronic tab of the CASTEP Calculation dialog) must be set to either
Schroedinger for non-relativistic treatment or ZORA for scalar-relativistic effects to be taken into
account. J-coupling calculations are supported only for systems with P1 symmetry. If your crystal has
symmetry other than P1, you will be prompted to convert the crystal symmetry to P1 before you can
continue.
G-tensor: Calculates the change in the EPR g-tensor for a molecule in a box or a defect in a crystal, as well
as the tensor of hyperfine parameters.
Note: The g-tensor and hyperfine parameters can only be calculated with spin-polarization turned on
and for systems with P1 symmetry. If your crystal has symmetry other than P1, you will be prompted
to convert the crystal symmetry to P1 before you continue.
System type: Specify the type of system used in the calculation. Options are:
Crystal
Molecule
Auto - CASTEP will assess the system and treat it as a crystal or molecule as appropriate
Note: Calculations in the Molecule mode can only be performed for systems with P1 symmetry. If
your crystal representing molecule in the box has symmetry other than P1, you will be prompted to
convert the crystal symmetry to P1 before you can continue.
More...: Provides access to the CASTEP NMR Options dialog, which allows you to specify the nuclear
quadrupolar moment for each element.
Note: Calculation of the NMR parameters requires higher accuracy for the wavefunctions than total
energy or geometry optimization calculations. This affects how you choose the Energy cutoff quality
parameter on the Electronic tab, even choosing Ultra-fine may not be sufficient to obtain high quality
NMR parameters. Therefore, it is advisable to study the convergence of the results as the Energy
cutoff parameter is increased. Generally, setting the cutoff to between 270 and 400 eV should
produce convergence of the shielding constant to about 1 ppm, but in some cases, the cutoff has to
be increased up to 550 eV.
Access methods
Note: NMR in CASTEP is part of the separately licensed module NMR CASTEP. NMR calculations can
only be performed if you have purchased this module.
Q moment: Specify the nuclear quadrupolar moment of the element displayed. Change the element
type using the dropdown list.
Tip: To reset the nuclear quadrupolar moment for the displayed element to the default value, enter 0
for the Q moment.
Reset all Q values to defaults: Sets all nuclear quadrupolar moments for all elements to the default
values.
Help: Displays the Help topic in a browser.
Access methods
k-point set: Specify the quality of the k-point set for the optical properties calculation. Each setting
corresponds to a particular separation between neighboring k-points in the Monkhorst-Pack grid.
Default = Medium.
More...: Provides access to the CASTEP Optical Properties Options dialog, which allows you to set
additional parameters, including the k-point set specification.
Access methods
Separation: When selected, generates the k-point grid according to the specified k-point separation.
Specify the k-point separation, in Å-1, in the associated text box.
Note: When you select the Separation, the Monkhorst-Pack parameters are derived to give the
specified separation between neighboring grid points.
Custom grid parameters: When selected, generates the k-point grid using the Monkhorst-Pack grid
parameters and the origin shift in fractional reciprocal space coordinates specified in the Grid
parameters and Origin shift text boxes, respectively.
Grid parameters: Specify the Monkhorst-Pack grid parameters in each of the lattice directions.
Actual spacing: Displays the k-point separation, in Å-1, resulting from the currently specified
Monkhorst-Pack grid parameters in each of the lattice directions.
Origin shift: Specify the offset of the Monkhorst-Pack grid in fractional reciprocal space coordinates.
Note: The Grid parameters and Origin shift controls are enabled only if you select the Custom grid
parameters option.
Display points...: Displays the number and fractional coordinates of the reciprocal space mesh points
that would be generated using the currently specified parameters.
Note: The actual set of k-points that will be used in the calculations may be altered if the symmetry of
the system changes.
Use separate XC functional for optics calculation: When checked, indicates that a different exchange-
correlation functional will be used for the optical properties calculation from the one used in the self-
consistent calculation.
Tip: This feature is particularly useful for carrying out expensive, screened exchange calculations.
Perform the main run with the LDA or GGA functional, then invoke screened or exact exchange in non
self-consistent mode during the band structure calculation.
Note: There is an important difference between standard DFT calculations with local exchange-
correlation potentials and the non-local exchange case. The potential used in the latter scenario
depends on the SCF k-points, while in the former case, the potential depends only on electron
density. This difference can make all properties calculations a lot more expensive in terms of memory
usage and CPU time, so it is advisable to limit the number of SCF k-points that you use in such
situations. This is particularly relevant for small unit cells, where the default settings may generate a
very large k-point set.
Functional: Select the type of DFT exchange-correlation potential to be used in the optical properties
calculation.
Choose the type of functional from the first dropdown list. Available options are:
LDA
GGA
HF
HF-LDA
sX
sX-LDA
PBE0
B3LYP
HSE03
HSE06
Then, if appropriate, select the specific functional from the second dropdown list. Depending on the
functional type selected from the first dropdown list, some of the following options may be available:
CA-PZ
PBE
RPBE
PW91
WC
PBESOL
BLYP
Note: The Functional control is enabled only if the Use separate XC functional for optics calculation
checkbox is checked.
Band energy tolerance: Specify the convergence criterion, in eV, for electronic eigenvalues during the
optical properties calculation. The calculation will stop when all electronic eigenvalues change by less
than this value over a single iteration of the electronic structure.
Help: Displays the Help topic in a browser.
Access methods
Orbitals selection
Checking the Orbitals checkbox on the Properties tab causes the orbitals (wavefunctions) to be
calculated as part of the CASTEP run.
Note: Orbitals are required to calculate the STM profile.
Access methods
Phonons selection
Selecting Phonons on the Properties tab displays options for calculating phonon properties.
Density of states: When selected, computes the phonon density of states as part of the CASTEP run.
Dispersion: When selected, computes the phonon dispersion as part of the CASTEP run.
Both: When selected, computes both the phonon density of states and dispersion during the CASTEP
run.
Calculate LO-TO splitting: When selected, applies a non-analytical correction to the dynamical matrix.
This calculates the LO-TO phonon frequency splitting at the Γ-point along the directions connecting the
Γ-point with the neighboring points along the reciprocal space path. Default = checked.
Method: Select the technique to use to calculate phonon frequencies. Available options are:
Linear response
Finite displacement
Finite displacement phonons are available only at the commensurate q-points (at the cost of creating a
supercell). However, in this case, you can obtain the phonon frequencies with no limitations on the
settings so that, for example, you can use ultrasoft potentials.
Note: If you request LO-TO splitting, then the restrictions on the system type and settings (insulator,
norm-conserving potentials) are the same for both the linear response and finite displacement
techniques.
More...: Provides access to the CASTEP Phonon Properties Setup dialog, where you can specify
additional parameters affecting a phonon calculation, including the convergence tolerance.
Access methods
Linear response
Use interpolation: When selected, calculates the real space dynamical matrix using the Monkhorst-Pack
mesh defined by the separation parameter.
q-vector grid spacing for interpolation: Specify the q-point separation parameter, in Å-1, which
represents the average distance between Monkhorst-Pack mesh q-points used in the real space
dynamical matrix calculations. A smaller separation parameter produces a denser mesh, giving more
accurate results at the expense of longer calculation times.
Adjust phonon q-vector grid: When selected, adjusts the phonon Monkhorst-Pack mesh to always
connect points on the electronic k-vector grid. This can significantly improve convergence especially for
metallic cases.
Note: You can only specify the q-vector grid spacing for interpolation and Adjust phonon q-vector
grid if you select Use interpolation.
Convergence tolerance: Specify the convergence criterion, in eV Å-2, for the force constants during a
phonon properties run.
Note: You can only specify the Convergence tolerance if you select Linear response as the Method.
Finite displacement
Use: Select the method to calculate the dynamic matrix using finite displacement. Available options are:
Several small supercells - uses the Monkhorst-Pack mesh defined by the separation parameter.
When you select Finite displacement as the Method, this invokes the nondiagonal supercell
approach of Lloyd-Williams and Monserrat (2015).
One large supercell
q-vector grid spacing for interpolation: Specify the q-point separation parameter, in Å-1. This
represents the average distance between Monkhorst-Pack mesh q-points used in the real space
dynamic matrix calculations. A smaller separation parameter produces a denser mesh, giving more
accurate results at the expense of longer calculation times.
Note: You can only specify the q-vector grid spacing for interpolation if you select Several small
supercells.
Supercell defined by cutoff radius: Specify the real space cutoff radius, in Å, for dynamical matrix
calculations. CASTEP uses this value to construct an appropriate supercell. A larger cutoff radius
produces more accurate results at the expense of longer calculation times.
If the cutoff radius specified creates a large supercell, its size is reported as a factor of the size of the
current cell, indicating that the calculation may be computationally expensive. If the cutoff radius is
reasonable, the report confirms that the original cell is large enough. For non-3D periodic systems, you
receive a reminder that Materials Studio cannot construct a supercell.
Note: You can only specify the Supercell defined by cutoff radius if you select One large supercell.
Dispersion
Quality: Specify the quality of the q-vector used for the phonon dispersion calculation. The quality level
affects the approximate separation between consecutive q-vectors on the reciprocal space path.
Available options and the corresponding q-vector separations are:
Separation: Specify the approximate separation between q-vectors in Å-1. The default depends on the
selected Quality of the q-vector.
Path...: Provides access to the Brillouin Zone Path dialog, which allows you to specify a custom
reciprocal space path.
Density of states
Quality: Specify the quality of the q-vector to use for the phonon density of states calculation. The
quality level affects the separation between neighboring q-vectors in the Monkhorst-Pack grid. Available
options and the corresponding q-vector separations are:
Note: For Linear response calculations, if you select Gamma point and select the Use interpolation,
this calculates the frequencies using the interpolation method. This involves calculations of phonon
frequencies at multiple q-vectors in the Brillouin zone and is therefore an expensive way of generating
Gamma point phonons. The only justification for such calculations is as a way of testing the accuracy
of the interpolation scheme (directly calculated frequencies are more accurate). To run calculations
only at the Gamma point, clear selection of Use interpolation.
More...: Provides access to the CASTEP Phonon Density of States Options dialog, where you can specify
the q-vector to use.
Help: Displays the Help topic in a browser.
Access methods
Separation: When selected, indicates that the q-vector grid will be generated according to the specified
q-vector separation. Specify the q-vector separation, in Å-1, in the associated text box.
Note: When the Separation option is selected, the Monkhorst-Pack parameters are derived so as to
produce the specified separation between neighboring grid points.
Custom grid parameters: When selected, indicates that the q-vector grid will be generated using the
Monkhorst-Pack grid parameters and the origin shift in fractional reciprocal space coordinates specified
in the Grid parameters and Origin shift text boxes, respectively.
Grid parameters: Specify the Monkhorst-Pack grid parameters in each of the lattice directions.
Actual spacing: Displays the k-point separation, in Å-1, resulting from the currently specified
Monkhorst-Pack grid parameters in each of the lattice directions.
Origin shift: Specify the offset of the Monkhorst-Pack grid in fractional reciprocal space coordinates.
Note: The Grid parameters and Origin shift controls are enabled only if you select the Custom grid
parameters option.
Display points...: Displays, in a new window, the number and fractional coordinates of the reciprocal
space mesh points that would be generated using the currently specified parameters.
Note: The actual set of q-vectors that will be used in the calculation may be altered if the symmetry of
the system changes.
Help: Displays the Help topic in a browser.
Access methods
Auto
Crystal
Molecule
Linear
Calculate Raman intensities: When checked, activities for Raman active vibrations will be calculated.
Subsequently CASTEP analysis can be used to construct Raman mode intensities for a given
temperature and laser wavelength from calculated activities.
More...: Provides access to the CASTEP IR and Raman Options dialog, which allows you to set additional
parameters affecting a polarizability, IR and Raman calculation, including the convergence tolerance and
Raman frequency range.
Access methods
Polarization selection
Checking the Polarization checkbox on the Properties tab includes calculation of polarization as part of
the CASTEP run. If the task is Elastic Constants, CASTEP uses calculated values of polarization to evaluate
the piezoelectric strain tensor.
Note: Polarization calculations are not allowed if you use any of the following settings:
Ultra-soft pseudopotentials
Spin-polarized calculation
DFT+U
Spin-orbit coupling with non-collinear spins
External electric field
Mixture atoms
Access methods
Access methods
Solvation energy
Checking the Solvation energy checkbox on the Properties tab causes calculation of the free energy of
solvation, as a separate CASTEP run. This additional run evaluates the energy of a molecule in a box both
in vacuum and in the presence of an implicit solvent. Then CASTEP produces the energy of solvation
from these results.
There are certain limitations for solvation energy calculations:
The system must represent a molecule in a box.
Symmetry of the system must be P1.
Calculations must not include spin-polarization.
Stress calculations are not available.
You can specify the options for the solvation energy calculation on the Solvent tab of the
CASTEP Electronic Options dialog.
Note: Use at least a Fine FFT grid Density for solvation calculations; the Precise density is even better.
You can specify the FFT grid Density on the Basis tab of the CASTEP Electronic Options dialog. Also use
an increased value of Augmentation density scaling factor, for example 1.5-2.0 instead of the default
value of 1.0.
An increased FFT density and energy cutoff can help to eliminate convergence errors for the multigrid
solver for electrostatics in a solvent.
Access methods
Stress selection
Checking the Stress checkbox on the Properties tab causes the stress tensor to be calculated as part of
the CASTEP run.
Access methods
Note: The options specified on the Job Control tab only apply to new jobs. They do not affect jobs
that are already running.
Gateway location: Select a server for the CASTEP calculation from the list of available server machines.
You can add servers to the list using the Server Console.
Queue: Specify the queue to which to submit the job. Select the required queue from the list, to display
the available queues on the chosen gateway. For more information, see Working with queues.
Note: For 3DEXPERIENCE Cloud jobs you cannot specify a Queue or choose a value for the Run in
parallel on number of cores. Instead, select the Category for the job, these options automatically
determine the number of cores to use for the job.
Category: For 3DEXPERIENCE jobs, select the compute power to use.
Job description: Specify the name to use to identify the job.
Materials Studio automatically assigns a default job description. You can specify an alternative
description by clearing the Automatic checkbox and entering the new name in the Job description text
box.
Note: Do not use this field for Properties task since in this case CASTEP Analysis will not be able to
associate results files with the original structure.
Automatic: Use the automatically generated job description. Default = checked.
Run in parallel on: Runs the job on the selected gateway using the specified number of computer cores.
The text to the right of the Run in parallel on control indicates the maximum number of cores available
on the selected gateway.
Note: It is recommended to use the number of cores which is commensurate with the number of k-
points in the calculation. You can determine the number of k-points by using the Display points...
button on the k-points tab of the CASTEP Electronic Options dialog.
Using up to: Specify the maximum number of threads allowed to be used by each MPI process. If the
number of threads is greater than 1, the hybrid OpenMP/MPI parallelization will be invoked. The actual
number of threads in use will depend on the details of the hardware configuration and the queuing
system setup. For example, under with some queuing systems you can set up the gateway to guarantee
exclusive use of nodes when number of threads is greater than 1.
Note: This functionality is available only on Linux gateways. To ensure it works as expected when a
gateway is using a queuing system, check that the queuing system is set up consistently. See Working
with queues for additional details.
Note: Hybrid OpenMP/MPI parallelization can substantially decrease the amount of memory
required for a job (though it could somewhat decrease speed).
Runtime optimization: Specify the strategy used to balance memory use and speed during the
calculation. Available options are:
Default - Balance speed and memory use. Wavefunction coefficients for all k-points in a calculation
will be kept in memory during, rather than be paged to disk. Some large work arrays will be paged to
disk.
Memory - Minimize memory use. All wavefunctions and large work arrays are paged to disk.
Speed - Maximize speed. No paging to disk.
Note: This option does not apply to calculations using the nonlocal exchange correlation potential. In
this case paging does not save any memory and Speed optimization is enforced.
More...: Provides access to the CASTEP Job Control Options dialog which allows you to set additional
options associated with monitoring and controlling the results of a CASTEP calculation.
Access methods
Tip: It may be useful to retain the server files so that standalone CASTEP runs can performed.
Automatically view output: When checked, the output files from the job will be opened automatically
when the calculation is completed. Files opened may include a structure document and a CASTEP
output (.castep) file for the main task performed by the job. Default = checked.
Notify on job completion: When checked, indicates that a dialog will be displayed when the job is
complete. Default = checked.
Tip: If you run several short jobs in one session, you may find it useful to stop the automatic display
of job completion notices and results files.
Return BibTeX list of references: When checked, indicates that a BibTeX formatted list of references
used for this run will be returned when the job has completed.
Return complete check file: When checked, indicates that the full check file will be returned when the
job has completes. Availability of this file (which can be very large) can accelerate the restart of a job.
Note: This option does not apply to calculations using the nonlocal exchange correlation potential. In
this case the check file is always returned so that restarts are enabled.
Help: Displays the Help topic in a browser.
Access methods
Access methods
Note: When the appropriate structure document is active and atoms are selected, the contribution
to the density of states from the selected atoms will be plotted. Otherwise, the contribution from all
atoms is considered.
Display DOS: Specify which spin components should be plotted in the density of states graph. For a
spin-polarized calculation, the supported options are:
Total - contributions from both spin-up and spin-down eigenstates are summed
Alpha - contributions from spin-up eigenstates only
Beta - contributions from spin-down eigenstates only
Alpha and Beta - contributions from both spin-up and spin-down eigenstates are displayed in a
butterfly plot
Spin - difference between contributions from spin-up and spin-down eigenstates is displayed
Total and Spin
When a non-spin-polarized CASTEP calculation is analyzed, only the Total option is available.
Function: Specify whether to plot the untreated DOS or the integrated density of states (the number of
states).
More...: Provides access to the CASTEP DOS Analysis Options dialog which allows you to specify
parameters controlling the density of states integration method.
Note: The Full, Partial, Display DOS, and More... options are enabled only if Show DOS is checked.
View: Displays the band structure using the options specified.
Note: If you have calculated DOS for a spin polarized system with a non-zero initial spin and the total
spin is not optimized two charts will automatically be displayed, <seedname> Band
Structure Alpha.xcd and <seedname> Band Structure Beta.xcd.
Access methods
Tip: If you included a core hole in your calculation, you can select the relevant atom using the Atom
Selection dialog.
Spectrum: Specify the type of spectrum that will be calculated. If there is a core hole in the shell a *
symbol is included in the spectrum name, for example 1s*[K1]. Available options (with and without core
holes) are:
1s[K1]
2s[L1]
2p[L23]
3s[M1]
3p[M23]
3d[M45]
4s[N1]
4p[N23]
4d[N45]
5s[O1]
5p[O23]
Type: Specify the type of spectrum that will be calculated. Available options are:
Absorption - the predicted absorption spectrum
Emission - the predicted emission spectrum
More...: Opens the CASTEP EELS Analysis Options dialog.
Calculation: Specify the polarization of the incident radiation for the core level spectroscopy calculation.
Polarization: Specify the direction of the polarization of the incident radiation in fractional (lattice)
coordinates. Not available when Unpolarized is selected for Calculation.
Atom Selection...: Opens the Atom Selection dialog.
View: Displays the selected core level spectrum.
A chart document is created using the information contained in the active grid document.
Note: If the requested spectrum is not available, a warning message will be returned explaining the
reasons.
Access methods
Lifetime effects: When checked, lifetime broadening effects will be taken into account.
Core level broadening: Specify the broadening to be used at the core level. Default values for lifetime
broadening are tabulated based on the compilation by Fuggle and Inglesfield (1992).
Reset broadening to default: Resets the Core level broadening to its default value.
Include spin-orbit splitting: When checked, spin-orbit splitting for the core level is added to the
spectrum.
Spin-orbit splitting: Specify the value of the core level spin-orbit splitting to be used. Default values for
spin-orbit splitting are taken from atomic spectroscopic data (see the WebElements Periodic Table:
http://www.webelements.com). These values represent the difference between reported binding
energies for core electrons of the given species, for example the difference between 2p 1/2 and 2p 3/2
gives the L2,3 splitting.
Reset spin-orbit splitting to default: Resets the Spin-orbit splitting to its default value.
Normalize spectrum: When checked, the resulting spectrum will be normalized to 1 at maximum.
Help: Displays the Help topic in a browser.
Access methods
Display DOS: Specify which spin components should be plotted in the density of states graph. For a
spin-polarized calculation, the supported options are:
Total - contributions from both spin-up and spin-down eigenstates are summed
Alpha - contributions from spin-up eigenstates only
Beta - contributions from spin-down eigenstates only
Alpha and Beta - contributions from both spin-up and spin-down eigenstates are displayed in a
butterfly plot
Spin - difference between contributions from spin-up and spin-down eigenstates is displayed
Total and Spin
When a non-spin-polarized CASTEP calculation is analyzed, only the Total option is available.
Note: If you have calculated DOS for a spin polarized system with a non-zero initial spin and the total
spin is not optimized only the Total option is available. In this case, two charts will automatically be
displayed, <seedname> DOS Alpha.xcd and <seedname> DOS Beta.xcd.
Function: Specify whether to plot the untreated DOS or the integrated density of states (the number of
states).
Scissors: Specify the scissors operator to be used in plotting the density of states.
The scissors operator is only applied for insulating systems, which have a clear separation between
valence band and conduction band states. It is ignored for metallic systems.
Atom Selection...: Opens the Atom Selection dialog.
More...: Provides access to the CASTEP DOS Analysis Options dialog, which allows you to specify
parameters controlling the density of states integration method.
View: Displays the density of states using the options specified.
Access methods
Note: This option is enabled only if the Smearing integration method is selected.
Accuracy level: Specify the quality of the interpolation to be used. Available options are:
Coarse - Interpolates onto a k-point grid of about 50×50×50
Medium - Interpolates onto a k-point grid of about 100×100×100
Fine - Interpolates onto a k-point grid of about 200×200×200
Instrument broadening: Specify an additional broadening to be applied to the interpolated density of
states.
Note: The Accuracy level and Instrument broadening options are enabled only if the Interpolation
integration method is selected.
Number of points per: Specify number of points per 1 eV or per 1 Ha energy range. Use high settings to
generate charts that remain smooth after a zoom-in operation.
OK: Updates the settings with any changes and closes the dialog.
Cancel: Closes the dialog without updating any settings.
Help: Displays the Help topic in a browser.
Access methods
Note: The empirical expressions for Vickers hardness and for fracture toughness cannot be applied if
the calculated shear modulus is negative. This can happen when the calculation settings are not
sufficiently accurate. In such a case, the summary report does not include the hardness or the
fracture toughness.
Access methods
Results file: Select the CASTEP results file from which the information required to visualize the electron
density difference will be taken. When more than one set of results is available, use the button to
browse the current directory and appropriate subdirectories for results files.
Use atomic densities: When checked, the electron density difference with respect to the sum of the
atomic densities will be created. If unchecked, the density difference with respect to the sum of the
densities of sets of atoms will be created.
View isosurface on import: When checked, an isosurface will be created automatically when the
electron density difference field is imported.
Import: Imports the selected electron density difference field from the CASTEP results file into the
current structure document.
Note: Analysis based on atomic densities requires the presence of a binary file with the extension
.chdiff; analysis based on sets of atoms requires that for each set of input files, seed_
Subset*.*, there are binary files with the extension .check present.
Access methods
Results file: Select the CASTEP results file from which the energy evolution information will be taken.
When more than one set of results is available, use the button to browse the current directory and
appropriate subdirectories for results files.
View: Generates chart documents as described in the Displaying trajectory and chart data topic.
These charts are the same as would be generated by selecting Update graphs in the CASTEP Job Control
Options dialog. Charts generated using Energy evolution can be generated whether or not you chose
Update graphs in the original calculation.
Note: Each time you select Update graphs the existing chart documents are rewritten, and any
changes or annotations you have made to the charts will be lost.
Access methods
IR spectrum selection
The following options are available when IR spectrum is selected from the list of properties.
Results file: Select the CASTEP results file from which the Hessian (dynamical matrix) and ATP tensor
(Born charges) information will be taken. When more than one set of results are available, use the
button to browse the current directory and appropriate subdirectories for results files.
Import Hessian from CASTEP output: Imports the Hessian and ATP tensor data, if present, from the
selected CASTEP results file into the current structure document.
Note: This only is enabled if the currently active document contains a structure for which CASTEP
results for the dynamical matrix at the Γ-point are available.
Once a Hessian has been imported, vibrational mode frequencies and, if ATP tensors were calculated,
absorption intensities can be displayed in the form of a list of values or graphically as a vibrational
spectrum using the Vibrational Analysis tool.
View vibrational analysis: Opens the Vibrational Analysis dialog. This is only enabled when the Hessian
is imported into the current structure document.
Access methods
NMR selection
The following options are available when NMR is selected from the list of properties.
Note: NMR in CASTEP is part of the separately licensed module NMR CASTEP. NMR calculations can
only be performed if you have purchased this module.
Results file: Select the CASTEP results file from which the NMR information will be taken. When more
than one set of results is available, use the button to browse the current directory and appropriate
subdirectories for results files.
Assign isotropic chemical shielding to structure: Adds the NMRShielding property to the atoms for
which NMR shielding data are available.
Assign electric field gradients to structure: Adds the EFGQuadrupolarCoupling and EFGAsymmetry
properties to the atoms for which electric field gradient data are available.
Once added, all three of these properties are accessible in the Properties Explorer and can be used for
labeling.
Show J-coupling constants: Generates and opens a study table reporting the calculated J-coupling
constants. This option is only available when the Results file selector is set to use a seedname_
NMRJC.castep file.
Access methods
Note: This button is enabled only if the current document is a grid document containing a dielectric
function.
A chart document is created using the dielectric function information contained in the active grid
document.
Access methods
Orbitals selection
The following options are available when Orbitals is selected from the list of properties.
Results file: Select the CASTEP results file from which the orbitals information will be taken. When more
than one set of results is available, use the button to browse the current directory and appropriate
subdirectories for results files.
Filter: Choose which orbitals are displayed in the table:
Alpha - displays data only for alpha-spin orbitals (spin up)
Beta - displays data only for beta-spin orbitals (spin down)
Note: For non-spin-polarized calculations, this option is not available.
The orbital analysis table, below the Filter dropdown list, lists the eigenvalues and other information
about each orbital. The list starts with the lowest energy orbital. Other information contained in the
table includes:
N - number of the orbital, with 1 corresponding to the lowest energy orbital
s - spin, + for alpha, and - for beta.
Note: In a non-spin-polarized calculation, all orbitals are labeled +.
From - lowest value of the energy of the Nth band at any k-point (in eV)
To - highest value of the energy of the Nth band at any k-point (in eV)
Scissors: Specify the scissors operator to be used in displaying the orbital eigenvalues.
The scissors operator is only applied for insulating systems, which have a clear separation between
valence band and conduction band states. It is ignored for metallic systems.
View isosurface on import: When checked, an isosurface will be created automatically when the orbital
field is imported. The units for the isovalues are e/Å3.
Import: Imports the selected orbital field from the CASTEP results file into the current structure
document.
Access methods
Note: When atoms are selected in the active structure document, the contribution to the phonon
density of states from only the selected atoms will be plotted. Otherwise, the contribution from all
atoms is considered.
More...: Provides access to the CASTEP Phonon DOS Analysis Options dialog which allows you to specify
parameters that control the density of states integration method and DOS normalization scheme.
Note: The DOS display and More options are enabled only if Show DOS is checked.
View: Displays the phonon dispersion using the options specified.
Access methods
Results file: Select the CASTEP results file from which the population analysis information will be taken.
When more than one set of results is available, use the button to browse the current directory and
appropriate subdirectories for results files.
Assign charges to structure: Imports the selected charges into the active structure document and
assigns them to each atom. You can see the charges by labeling the atoms according to their charge.
Select the type of charges that you wish to import from the dropdown list, available options are:
Mulliken
Hirshfeld
Assign spins to structure: Imports the selected spins into the active structure document and assigns
them to each atom. You can see the spins by labeling the atoms according to their spin. Select the type
of spins that you wish to import from the dropdown list, available options are:
Mulliken
Hirshfeld
Assign Mulliken bond orders to structure: Imports Mulliken bond orders into the active structure
document and assigns them to each bond.
Access methods
Potentials selection
The following options are available when Potentials is selected from the list of properties.
Results file: Select the CASTEP results file from which the potentials information will be taken. When
more than one set of results is available, use the button to browse the current directory and
appropriate subdirectories for results files.
Note: The local potential cannot be analyzed for calculations with spin-orbit interaction taken into
account.
Potential field: Indicates the type of potential field to be imported. Currently only SCF local potential is
available.
View isosurface on import: When checked, an isosurface will be created automatically when the
potential field is imported.
Import: Imports the selected potential field from the CASTEP results file into the current structure
document. If the structure is a slab with a region of vacuum, a Chart document named <seedname>
Potential Profile.xcd is also displayed containing a plot of the potential averaged in the
planes perpendicular to the surface normal. This chart document also contains the value of work
function calculated as a difference between the potential level in a vacuum and the Fermi energy.
Note: When calculating a work function from a periodic slab model, you should ensure that the self-
consistent dipole correction is activated on the SCF tab of the CASTEP Electronic Options dialog.
Access methods
Note: In some reactions, the transition state may have optical isomers with degeneracy n. These
require the prefactor of the Arrhenius rate to be corrected by a factor n. Materials Studio does not
automatically include this prefactor, users are encouraged to add it to the results study table by
hand.
Calculate: Calculate and display a study table containing rate coefficients for the forward and reverse
reactions.
Note: The prefactor for the modified Arrhenius equation is calculated using T0= 298.15 K as reference
temperature.
Access methods
Results file: Select the CASTEP results file from which the Hessian (dynamical matrix) and ATP tensor
(Born charges) information will be taken. When more than one set of results are available, use the
button to browse the current directory and appropriate subdirectories for results files.
Function: Specify whether to calculate the Intensity or the Activity of the vibrational mode.
Temperature: If the intensity of the vibrational mode is being calculated set the temperature in Kelvin.
Incident light: If the intensity of the vibrational mode is being calculated specify the wavelength of the
incident light in nm. The default value of 514.5 nm corresponds to a standard Ar laser. Other popular
wavelengths for laser Raman spectrometers include 632.8 nm (He-Ne), 785 nm (diode), and 1064 nm
(Nd:YAG).
Smearing: Specify the type of broadening (Gaussian or Lorentzian) and width to be used. Default
Lorentzian broadening with width = 20.0 cm-1.
Units: Specify the units for the X axis of the spectrum, options are:
meV
THz
cm-1
Reverse wavenumber axis: When checked, indicates that the values on the X axis will decrease from
right to left.
Reverse intensity axis: When checked, indicates that the values on the Y axis will decrease from bottom
to top.
Once a Hessian has been imported, vibrational mode frequencies and, if ATP tensors were calculated,
absorption intensities can be displayed in the form of a list of values or graphically as a vibrational
spectrum using the Vibrational Analysis tool.
View: Displays the selected Raman spectrum.
Access methods
Note: When more than one set of results is available, it is preferable to use output files for DOS
calculation, especially when STM images of empty bands are required.
STM bias: Specify the value of the bias for STM modeling. Positive bias corresponds to the imaging of
empty (conduction) states, negative bias to the imaging of occupied (valence) states.
View isosurface on import: When checked, an isosurface will be created automatically when the STM
profile field is imported.
Import: Imports the selected STM profile field from the CASTEP results file into the current structure
document.
Access methods
Structure selection
The following options are available when Structure is selected from the list of properties.
Results file: Select the CASTEP results file from which the structure information will be taken. When
more than one set of results is available, use the button to browse the current directory and
appropriate subdirectories for results files.
Update structure from CASTEP output: Restores atomic coordinates and cell parameters from the
CASTEP results file.
Note: This button is enabled only if the current document contains a structure for which CASTEP
results are available.
Create trajectory: Creates a trajectory (.xtd) file from the results of a CASTEP Geometry Optimization,
Dynamics or TS Search task.
Note: This button is enabled only if the current document contains a structure for which CASTEP
trajectory results are available.
Access methods
Tip: The lower limit should be no less than be 5 K to avoid numerical noise due to inadequate Brillouin
zone sampling at low temperatures. The upper limit should not be significantly greater than the
Debye temperature because of the limited accuracy of the quasiharmonic approximation at high
temperatures.
Plot Debye temperature: When checked, indicates that a separate chart with the temperature
dependence of the Debye temperature will be displayed.
Assign temperature factors to structure: Adds temperature factor information to the structure.
View: Displays the thermodynamics properties using the options specified.
Note: Thermodynamic calculations performed only at the Γ-point give very poor results, as no
acoustic degrees of freedom can be taken into account.
Note: Heat capacity is measured in calories per unit cell per Kelvin. The unit cell can contain several
formula units.
Access methods
Dziedzic, J.; Helal, H. H.; Skylaris, C.-K.; Mostofi, A. A.; Payne, M. C. "Minimal parameter implicit solvent
model for ab initio electronic-structure calculations", Europhys. Lett., 95, 43001 (2011).
Dziedzic, J.; Fox, S. J.; Fox, T.; Tautermann, C. S.; Skylaris, C.-K. "Large-Scale DFT Calculations in Implicit
Solvent - A Case Study on the T4 Lysozyme L99A/M102Q Protein", Int. J. Quantum Chem., 113, 771-785
(2013).
Fattebert, J.-L.; Gygi, F. "Density Functional Theory for Efficient Ab Initio Molecular Dynamics
Simulations in Solution", J. Comput. Chem., 23, 662-666 (2002).
Fattebert, J.-L.; Gygi, F. "First-Principles Molecular Dynamics Simulations in a Continuum Solvent", Int. J.
Quant. Chem., 93, 139-147 (2003).
Flubacher, P.; Leadbetter, A. J.; Morrison, J. A. "The heat capacity of pure silicon and germanium and
properties of their vibrational frequency spectra", Phil. Mag., 4, 273-294 (1959).
Fox, S. J.; Dziedzic, J.; Fox, T.; Tautermann, C. S.; Skylaris, C.-K. "Density functional theory calculations
on entire proteins for free energies of binding: application to a model polar binding site", Proteins, 82,
3335-3346 (2014).
Francis, G. P.; Payne, M. C. "Finite Basis Set Corrections to Total Energy Pseudopotential Calculations",
J. Phys.: Condens. Matter, 2, 4395-4404 (1990).
Frantti, J. et al. "Neutron diffraction studies of Pb(ZrxTi1-x)O3 ceramics", Jpn. J. Appl. Phys., Pt. I, 39,
5697-5703 (2000).
Fuggle, J.C., Inglesfield, J.E. (Eds.) "Unoccupied Electronic States", Springer, Berlin (1992).
Gale, J. D.; Rohl, A. L.; Milman, V.; Warren, M. C. "An ab initio study of the structure and properties of
aluminium hydroxide: gibbsite and bayerite", J. Phys. Chem. B, 105, 10236-10242 (2001).
Gao, S-P.; Pickard, C.J.; Payne, M.C.; Zhu, J.; Yuan, J. "Theory of core-hole effects in 1s core-level
spectroscopy of the first-row elements", Phys. Rev. B, 77, 115122 (2008).
Godby, R. W. "Exchange and correlation in solids", Top. Appl. Phys., 69, 51-88 (1992).
Gonze, X.; Allan, D. C.; Teter, M. P. "Dielectric tensor, effective charges, and phonons in a-quartz by
variational density-functional perturbation theory", Phys. Rev. Lett., 68, 3603-3606 (1994).
Gonze, X. "First-principles responses of solids to atomic displacements and homogeneous electric fields:
Implementation of a conjugate-gradient algorithm", Phys. Rev. B, 55, 10337-10354 (1997).
Govind, N.; Petersen, M.; Fitzgerald, G.; King-Smith, D.; Andzelm, J. "A generalized synchronous transit
method for transition state location", Comput. Mater. Sci., 28, 250-258 (2003).
Green, T. F. G.; Yates, J. R. "Relativistic nuclear magnetic resonance J-coupling with ultrasoft
pseudopotentials and the zeroth-order regular approximation", J. Chem. Phys., 140, 234106 (2014).
Hakala, M.; Puska, M. J.; Nieminen, R. M. "First-principles calculations of interstitial boron in silicon",
Phys. Rev. B, 61, 8155-8161 (2000).
Hamann, D. R.; Schluter, M.; Chiang, C. "Norm-Conserving Pseudopotentials", Phys. Rev. Lett., 43, 1494-
1497 (1979).
Hammer, B.; Hansen, L. B.; Norskov, J. K. "Improved adsorption energetics within density-functional
theory using revised Perdew-Burke-Ernzerhof functionals", Phys. Rev. B, 59, 7413-7421 (1999).
Harris, R.K. "Nuclear spin properties and notation", In: Encyclopedia of nuclear magnetic resonance,
Grant, D.M.; Harris, R.K., editors. Chichester: John Wiley & Sons; 5, 3301 (1996).
Hebert, C. "Practical aspects of running the WIEN2k code for electron spectroscopy", Micron, 38, 12-28
(2007).
Henkelman, G.; Jónsson, H. "Improved tangent estimate in the nudged elastic band method for finding
energy paths and saddle points", J. Chem. Phys., 113, 9978 (2000).
Heßelman, A.; Görling, A. "Blindness of the Exact Density Response Function to Certain Types of
Electronic Excitations: Implications for Time-Dependent Density-Functional Theory", Phys. Rev. Lett.,
102, 233003 (2009).
Heyd, J.; Scuseria, G. E.; Ernzerhof, M. "Hybrid functionals based on a screened Coulomb potential", J.
Chem. Phys., 118, 8207-8215 (2003).
Hirata, S.; Head-Gordon, M. "Time-dependent density functional theory within the Tamm-Dancoff
approximation", Chem. Phys. Lett., 314, 291-299 (1999).
Hirshfeld, F. L. "Bonded-Atom Fragments for Describing Molecular Charge Densities", Theoret. Chim.
Acta, 44, 129-138 (1977).
Ho, K. M.; Fu, C. L.; Harmon, B. N. "Vibrational frequencies via total energy calculations - Applications to
transition metals", Phys. Rev. B, 29, 1575-1587 (1984).
Hutter, J. "Excited state nuclear forces from the Tamm-Dancoff approximation to time-dependent
density functional theory within the plane wave basis set framework", J. Chem. Phys., 118, 3928-3934
(2003).
Iniguez, J.; Vanderbilt, D.; Bellaiche, L. "First-principles study of (BiScO3)1-x-(PbTiO3)x piezoelectric
alloys", Phys. Rev. B, 67, 224-107 (2003).
IUPAC, "Notations and conventions in molecular spectroscopy: Part 2. Symmetry notation.", Pure and
Appl. Chem., 69 , 1641-1649 (1997).
Jirak, Z.; Kala, T. "Crystal Structures of Ferroelectric Phases FR(LT)/FR(HT) In PbZr0.75Ti0.25O3 Solid
Solutions and Their Dependence on Temperature", Ferroelectrics, 82, 79-84 (1988).
Joyce, S. A.; Yates, J. R.; Pickard, C. J.; Mauri, F. "A first principles theory of nuclear magnetic resonance -
coupling in solid-state systems", J. Chem. Phys., 127, 204107 (2007).
Joyce, S. A.; Yates, J. R.; Pickard, C. J.; Brown, S. P. "Density Functional Theory Calculations of Hydrogen-
Bond-Mediated NMR J Coupling in the Solid State", J. Am. Chem. Soc., 130, 12663-12670 (2008).
Juan, Y. M.; Kaxiras, E.; Gordon, R. G. "Use of the generalized gradient approximation in
pseudopotential calculations of solids", Phys. Rev. B, 51, 9521-9525 (1995).
Karki, B. B.; Stixrude, L.; Clark, S. J.; Warren, M. C.; Ackland, G. J.; Crain, J. "Structure and elasticity of
MgO at high pressure", Am. Mineral., 82, 51-60 (1997).
Karki, B. B.; Stixrude, L.; Clark, S. J.; Warren, M. C.; Ackland, G. J.; Crain, J. "Elastic properties of
orthorhombic MgSiO3 perovskite at lower mantle pressures", Am. Mineral., 82, 635-638 (1997).
Karki, B. B.; Stixrude, L.; Crain, J. "Ab initio elasticity of three high-pressure polymorphs of silica",
Geophys. Res. Lett., 24, 3269-3272 (1997).
Karki, B. B.; Clark, S. J.; Warren, M. C.; Hsueh, H. C.; Ackland, G. J.; Crain, J. "Ab initio elasticity and lattice
dynamics of AgGaSe2", J. Phys.: Cond. Matt., 9, 375-380 (1997).
Kerker, G. "Non-singular atomic pseudopotentials for solid state applications", J. Phys. C, 13, L189-L194
(1980).
Kern, G.; Kresse, G.; Hafner, J. "Ab initio calculation of the lattice dynamics and phase diagram of boron
nitride", Phys. Rev. B, 59, 8551-8559 (1999).
King-Smith, R. D.; Payne, M. C.; Lin, J. S. "Real-space implementation of nonlocal pseudopotentials for
first-principles total-energy calculations", Phys. Rev. B, 44, 13063-13066 (1991).
Kleinman, L.; Bylander, D. M. "Efficacious Form for Model Pseudopotentials", Phys. Rev. Lett., 48, 1425-
1428 (1982).
Koelling, D. D.; Harmon, B. N. "A technique for relativistic spin-polarised calculations", J. Phys. C: Solid
State Phys., 10, 3107-3114 (1977).
Kresse, G.; Furthmuller, J. "Efficient iterative schemes for ab initio total-energy calculations using a
plane-wave basis set", Phys. Rev. B, 54, 11169-11186 (1996).
Krukau, A. V.; Vydrov, O. A.; Izmaylov, A. F.; Scuseria, G. E. "Influence of the exchange screening
parameter on the performance of screened hybrid functionals", J. Chem. Phys., 125, 224106 (2006).
Kunc, K.; Needs, R. J.; Nielsen, O.H.; Martin, R.M. Symmetry and Special Points Program K290 (1985).
Laasonen, K.; Pasquarello, A.; Car, R.; Lee, C.; Vanderbilt, D. "Car-Parrinello molecular dynamics with
Vanderbilt ultrasoft pseudopotentials", Phys. Rev. B, 47, 10142-10153 (1993).
Lee, C.; Yang, W.; Parr, R.G. "Development of the Colle-Salvetti correlation-energy formula into a
functional of the electron density", Phys. Rev. B, 37, 785-789 (1988).
Lee, M.H. PhD Thesis, Cambridge University (1996).
Leimkuhler, B.; Noorizadeh, E.; Theil, F. "A Gentle Stochastic Thermostat for Molecular Dynamics", J.
Stat. Phys., 135, 261-277 (2009).
Lejaeghere, K.; Van Speybroeck, V.; Van Oost, G.; Cottenier, S. "Error Estimates for Solid-State Density-
Functional Theory Predictions: An Overview by Means of the Ground-State Elemental Crystals", Critical
Reviews in Solid State and Materials Sciences, 39, 1-24 (2014).
van Lenthe, E.; Baerends, E. J.; Snijders, J. G. "Relativistic total energy using regular approximations", J.
Chem. Phys. , 101, 9783-9792 (1994).
Lin, J. S.; Qteish, A.; Payne, M. C.; Heine, V. "Optimized and transferable nonlocal separable ab initio
pseudopotentials", Phys. Rev. B, 47, 4174-4180 (1993).
Lin, J.; Lee, M. H.; Liu, Z. P.; Chen, C. T.; Pickard, C. J. "Mechanism for linear and nonlinear optical effects
in β-BaB2O4 crystals", Phys. Rev. B, 60, 13380-13389 (1999) .
Lindh, R.; Bernhardsson, A.; Karlström, G.; Malmqvist, P.-A. "On the use of a Hessian model function in
molecular geometry optimizations", Chemical Physics Letters, 241, 423-428 (1995).
Lloyd-Williams, J. H.; Monserrat, B. "Lattice dynamics and electron-phonon coupling calculations using
nondiagonal supercells", Phys. Rev. B, 92, 184301 (2015).
Louie, S. G.; Froyen, S.; Cohen, M. L. "Nonlinear ionic pseudopotentials in spin-density-functional
calculations", Phys. Rev. B, 26, 1738-1742 (1982).
Makov, G.; Payne, M. C. "Periodic boundary conditions in ab initio calculations", Phys. Rev. B, 51, 4014-
4022 (1995).
Marlo, M.; Milman, V. "Density-functional study of bulk and surface properties of titanium nitride using
different exchange-correlation functionals", Phys. Rev. B, 62, 2899-2907 (2000).
Marzari, N.; Vanderbilt, D.; Payne, M. C. "Ensemble density-functional theory for ab initio molecular
dynamics of metals and finite-temperature insulators", Phys. Rev. Lett., 79, 1337-1340 (1997).
Mazhnik, E.; Oganov, A. R. "A model of hardness and fracture toughness of solids", J. Appl. Phys., 126,
125109 (2019).
Mazzi, F.; Galli, E.; Gottardi, G. "The crystal structure of tetragonal leucite", Am. Miner., 61, 108-115
(1976).
Milman, V.; Payne, M. C.; Heine, V.; Needs, R. J.; Lin, J. S. "Free energy and entropy of diffusion by ab
initio molecular dynamics: Alkali ions in silicon", Phys. Rev. Lett., 70, 2928-2931 (1993).
Milman, V.; Lee, M. H.; Payne, M. C. "Ground-state properties of CoSi2 determined by a total-energy
pseudopotential method", Phys. Rev. B, 49, 16300 (1994).
Milman, V.; Winkler, B.; White, J. A.; Pickard, C. J.; Payne, M. C.; Akhmatskaya, E. V.; Nobes, R. H.
"Electronic structure, properties and phase stability of inorganic crystals: A pseudopotential plane-wave
study", Int. J. Quantum Chem., 77, 895-910 (2000).
Milman, V.; Warren, M. C. "Elasticity of hexagonal BeO", J. Phys.: Condens. Matter, 13, 241-245 (2001).
Milman, V.; Warren, M. C. "Elastic properties of TiB2 and MgB2", J. Phys.: Condens. Matter, 13, 5585-
5595 (2001b).
Milman, V. et al. "Electron and vibrational spectroscopies using DFT, plane waves and pseudopotentials:
CASTEP implementation", Journal of Molecular Structure: THEOCHEM, 954, 22-35 (2010).
Minkiewicz, V. J.; Shirane, G.; Nathans, R. "Phonon Dispersion Relation for Iron", Phys. Rev., 162, 528-
531 (1967).
Miwa, K. "Prediction of Raman spectra with ultrasoft pseudopotentials", Phys. Rev. B, 84, 094304 (2011).
Mohr, P. J.; Taylor, B. N.; Newell, B. D. "CODATA recommended values of the fundamental physical
constants: 2006", Rev. Mod. Phys., 80, 633-730 (2008).
Monkhorst, H. J.; Pack, J. D. "Special points for Brillouin-zone integrations", Phys. Rev. B, 13, 5188-5192
(1976).
Monkhorst, H. J.; Pack, J. D. "Special points for Brillouin-zone integrations" - a reply", Phys. Rev. B, 16,
1748-1749 (1977).
Montanari, B.; Harrison, N. M. "Lattice dynamics of TiO2 rutile: influence of gradient corrections in
density functional calculations", Chem. Phys. Lett., 364, 528-534 (2002).
Mulliken, R. S. J. Chem. Phys., 23, 1833-1846 (1955).
Neugebauer, J.; Scheffler, M. "Adsorbate-substrate and adsorbate-adsorbate interactions of Na and K
adlayers on Al(111)", Phys. Rev. B, 46, 16067-16080 (1992).
Niklasson, A. M. N.; Steneteg, P.; Bock, N. "Extended Lagrangian free energy molecular dynamics", J.
Chem. Phys., 135, 164111 (2011).
Niu, H.; Niu, S.; Oganov, A. R. "Simple and accurate model of fracture toughness of solids", J. Appl.
Phys., 125, 065105 (2019).
Nosé, S. "A molecular dynamics method for simulations in the canonical ensemble", Mol. Phys., 52, 255-
268 (1984).
Nye, J. F. Physical Properties of Crystals, Clarendon: Oxford (1957).
Packwood, D. et al. "A universal preconditioner for simulating condensed phase materials", J. Chem.
Phys., 144, 164109 (2016).
Palik, E. D. Handbook of optical constants of solids, Academic Press: Orlando (1985).
Parrinello, M; Rahman, A. J. Phys. (Paris), C6, 511 (1981).
Parlinski, K.; Li, Z. Q.; Kawazoe, Y. "First-Principles Determination of the Soft Mode in Cubic ZrO2", Phys.
Rev. Lett., 78, 4063-4066 (1997).
Payne, M.C.; Teter, M. P.; Allan, D. C.; Arias, T. A.; Joannopoulos, J. D. "Iterative Minimization
Techniques for Ab Initio Total Energy Calculations: Molecular Dynamics and Conjugate Gradients", Rev.
Mod. Phys., 64, 1045-1097 (1992).
Perdew, J. P.; Zunger, A. "Self-interaction correction to density-functional approximations for many-
electron systems", Phys. Rev. B, 23, 5048-5079 (1981).
Perdew, J. P.; Chevary, J. A.; Vosko, S. H.; Jackson, K. A.; Pederson, M. R.; Singh, D. J.; Fiolhais, C.
"Atoms, molecules, solids, and surfaces: Applications of the generalized gradient approximation for
exchange and correlation", Phys. Rev. B, 46, 6671-6687 (1992).
Perdew, J.P.; Burke, K.; Ernzerhof, M. "Generalized Gradient Approximation Made Simple", Phys. Rev.
Lett., 77, 3865-3868 (1996).
Perdew, J.P. et al. "Restoring the Density-Gradient Expansion for Exchange in Solids and Surfaces", Phys.
Rev. Lett., 100, 136406 (2008).
Pfrommer, B. G.; Cote, M.; Louie S. G.; Cohen, M. L. "Relaxation of Crystals with the Quasi-Newton
Method", J. Comput. Phys., 131, 233-240 (1997).
Pickard, C. J.; Mauri, F. "All-electron magnetic response with pseudopotentials: NMR chemical shifts",
Phys. Rev. B, 63, 245101 (2001).
Pickard, C. J.; Mauri, F. "First-Principles Theory of the EPR g Tensor in Solids: Defects in Quartz", Phys.
Rev. Lett., 88, 086403 (2002).
Pietrucci, F.; Bernasconi, M.; Di Valentin, C.; Mauri, F.; Pickard, C. J. "EPR g-tensor of paramagnetic
centers in yttria-stabilized zirconia from first-principles calculations", Phys. Rev. B, 73, 134112 (2006).
Porezag, D.; Pederson M. R. "Infrared intensities and Raman-scattering activities within density-
functional theory", Phys. Rev. B, 54, 7830 (1996).
Probert, M. I. J. "Improved algorithm for geometry optimization using damped molecular dynamics", J.
Comput. Phys., 191, 130-146 (2003).
Profeta, M.; Benoit, M.; Mauri, F.; Pickard, C. J. "First-Principles Calculation of the 17O NMR Parameters
in Ca Oxide and Ca Aluminosilicates: the Partially Covalent Nature of the Ca-O Bond, a Challenge for
Density Functional Theory", J. Am. Chem. Soc., 126, 12628-12635 (2004).
Puska, M. J. "Point defects in silicon, first-principles calculations", Comput. Mater. Sci., 17, 365-373
(2000).
Pyykkö, P. "Year-2008 nuclear quadrupole moments", Molecular Physics, 106, 1965-1974 (2008).
Ramer, N. J.; Rappe, A. M. "Virtual-crystal approximation that works: Locating a compositional phase
boundary in Pb(Zr1-xTix)O3", Phys. Rev. B, 62, 743-746 (2000).
Ranganathan, S. I.; Ostoja-Starzewski, M. "Universal Elastic Anisotropy Index", Phys. Rev. Lett., 101,
055504 (2008).
Rappe, A. M.; Rabe, K. M.; Kaxiras, E.; Joannopoulos, J. D. "Optimized pseudopotentials", Phys. Rev. B,
41, 1227-1230 (1990).
Read, A. J.; Needs, R. J. "Calculation of optical matrix elements with nonlocal pseudopotentials", Phys.
Rev. B, 44, 13071-13073 (1991).
Sahl, K. "Refinement of the crystal structure of bicchulite Ca2[Al2SiO6](OH)2", Z. Kristallogr., 152, 13-21
(1980).
Sanchez-Portal, D.; Artacho, E.; Soler, J. M. "Projection of plane-wave calculations into atomic orbitals",
Solid State Commun., 95, 685-690 (1995).
Savin, A.; Jepsen, O.; Flad, J.; Andersen, O. K.; Preuss, H.; von Schnering, H. G. "Electron localization in
solid-state structures of the elements - the diamond structure", Angewandte Chemie-International
Edition , 31, 187-188 (1992).
Scherlis, D. A.; Fattebert, J.-L.; Gygi, F.; Cococcioni, M.; Marzari, N. "A unified electrostatic and cavitation
model for first-principles molecular dynamics in solution", J. Chem. Phys., 124, 074103 (2006).
Segall, M. D.; Pickard, C. J.; Shah, R.; Payne, M. C. "Population analysis in plane wave electronic
structure calculations", Mol. Phys., 89, 571-577 (1996).
Segall, M. D.; Shah, R.; Pickard, C. J.; Payne, M. C. "Population analysis of plane-wave electronic
structure calculations of bulk materials", Phys. Rev. B, 54, 16317-16320 (1996).
Segall, M. D.; Lindan, P. J. D.; Probert, M. J.; Pickard, C. J.; Hasnip, P. J.; Clark, S. J.; Payne, M. C. "First-
principles simulation: ideas, illustrations and the CASTEP code", J. Phys.: Condens. Matter, 14, 2717-
2743 (2002).
Seidl, A.; Gorling, A.; Vogl, P.; Majewski, J. A.; Levy, M."Generalized Kohn-Sham schemes and the band-
gap problem", Phys. Rev. B, 53, 3764-3774 (1996).
Shih, B-C.; Yates , J.R. "Gauge-including projector augmented-wave NMR chemical shift calculations with
DFT+U", Phys. Rev. B, 96, 045142 (2017).
Spaldin, N.A. "A beginner's guide to the modern theory of polarization", J. Solid State Chem., 195, 2-10
(2012).
Srivastava, G. P.; Weaire, D. "The theory of the cohesive energies of solids", Adv. Phys., 26, 463-517
(1987).
Steneteg, P.; Abrikosov, I.A.; Weber, V.; Niklasson, A.M.N. "Wavefunction extended Lagrangian Born-
Oppenheimer molecular dynamics", Phys. Rev. B, 92, 075110 (2010).
Sun, J.; Remsing, R. C.; Zhang, Y. ; Sun, Z.; Ruzsinszky, A.; Peng ,H.; Yang, Z. ; Paul, A. ; Waghmare, U.
;Wu, X. ;Klein, M. L. ; Perdew, J. P. Nat. Chem. 8 , 831 (2016).
Tersoff, J.; Hamman, D. R. "Theory of the scanning tunneling microscope", Phys. Rev. B, 31, 805-813
(1985).
Tian, Y.; Xu, B.; Zhao, Z. "Microscopic theory of hardness and design of novel superhard crystals", Int. J.
Refr. Met. Hard Mater., 33, 93-106 (2012).
Troullier, N.; Martins, J. L. "Efficient pseudopotentials for plane-wave calculations", Phys. Rev. B, 43,
1993-2006 (1991).
Tuckerman, M. E.; Liu, Y.; Ciccotti, G.; Martyna, G. J. "Non-Hamiltonian molecular dynamics:
Generalizing Hamiltonian phase space principles to non-Hamiltonian systems", J. Chem. Phys., 115,
1678-1702 (2001).
Vanderbilt, D. "Soft self-consistent pseudopotentials in a generalized eigenvalue formalism", Phys. Rev.
B, 41, 7892-7895 (1990).
Walker, D. et al. "Halite-sylvite thermoelasticity", Am. Mineral., 89, 204-210 (2004).
Warren, J. L.; Worlton, T. G. "Improved version of group-theoretical analysis of lattice dynamics", Comp.
Phys. Commun., 8, 71-84 (1974).
Warren, M. C.; Ackland, G. J. "Ab Initio Studies of Structural Instabilities in Magnesium Silicate
Perovskite", Phys. Chem. Miner., 23, 107-118 (1996).
Warren, M. C.; Ackland, G. J.; Karki, B. B.; Clark, S. J. "Phase Transitions in Silicate Perovskites from First
Principles", Mineral. Mag., 62, 585-598 (1998).
Wei, S. Q.; Chou, M. Y. "Phonon dispersions in silicon and germanium from first principles calculations",
Phys. Rev. B, 50, 4859-4862 (1994).
White, J. A.; Bird, D. M. "Implementation of gradient-corrected exchange-correlation potentials in Car-
Parrinello total-energy calculations", Phys. Rev. B, 50, 4954-4957 (1994).
Winkler, B.; Dove, M. T.; Leslie, M. "Static lattice energy minimization and lattice dynamics calculations
on aluminosilicate minerals", Am. Mineral., 76, 313-331 (1991).
Winkler, B.; Milman, V. "Polymorphism of CsI", J. Phys.: Condens. Matter, 9, 9811-9817 (1997).
Winkler, B. "An introduction to 'Computational Crystallography'", Z. Kristallogr., 214, 506-527 (1999).
Winkler, B.; Pickard, C. J.; Segall, M. D.; Milman, V. "Density functional study of charge ordering in Cs2Au
(I)Au(III)Cl6 under pressure", Phys. Rev. B, 63, 14103 (2001).
Winkler, B.; Hytha, M.; Warren, M. C.; Milman, V.; Gale, J.; Shreuer, J. "Calculation of the elastic
constants of the Al2SiO5 polymorphs andalusite, sillimanite and kyanite", Z. Kristallogr., 216, 67-70
(2001).
Winkler, B.; Pickard, C. J.; Milman, V. "Applicability of a quantum mechanical "virtual crystal
approximation" to study Al/Si-disorder", Chem. Phys. Lett., 362, 266-270 (2002).
Wu, Z.; Cohen, R. E. "More accurate generalized gradient approximation for solids", Phys. Rev. B, 73,
235116 (2006).
Yanson, I. K. "Point contact spectroscopy of metals", Phys. Script., T23, 88-94 (1988).
Yates, J. R.; Pickard, C. J.; Mauri, F. "Calculation of NMR chemical shifts for extended systems using
ultrasoft pseudopotentials", Phys. Rev. B, 76, 024401 (2007).
Yates, J. R."Prediction of NMR J-coupling in solids with the planewave pseudopotential approach",
Magnetic Resonance in Chemistry, 48, S23-S31 (2010).
Yeh, I-C; Berkowitz, M.L. "Ewald summation for systems with slab geometry", J. Chem Phys., 111, 3155-
3162 (1999).
Yin, M. T.; Cohen, M. L. "Theory of lattice dynamical properties of solids - Application to Si and Ge",
Phys. Rev. B, 26, 3259-3272 (1982).